JPH04207352A - Encoding device and encoding method for image data - Google Patents

Abstract

PURPOSE:To facilitate coding in such a manner that a code quantity is confined within the specified code quantity within the specified processing time without impairing image quality by providing code quantity calculating means, reference code determining means, code quantity allotting means, and variable length encoding discontinuing means for respective blocks. CONSTITUTION:The statistical processing to obtain the code quantity for each of the respective blocks is executed by quantization 6 at an initially provisionally determined quantization width 12 and variable length encoding 8. The code quantity for each of the respective blocks determined by this statistical processing, the optimum quantization width 12 suitable from the target value of the total code quantity per image, and the allotted code quantity 20 for each of the respective blocks are then determined. The code quantity of each block by each color component is calculated in the variable length encoding calculating means 14. The target value of the total code quantity is proportionally distributed and the reference code quantity of each block by each color component is determined in the determining means 20. The allotting means 20 discontinues the variable length encoding so as not to exceed the allotted code quantity at the time of encoding processing. The encoding processing is so executed that the code quantity is confined within the specified code quantity in the specified processing time without impairing the image quality in this way.

Description

[Detailed description of the invention]

[Industrial Application Field] The present invention relates to an encoding device and encoding method for highly compression encoding image data to a predetermined capacity. [Prior Art] When image signals captured by a solid-state imaging device such as a CCD are stored as digital data in a storage device such as a memory cart, magnetic disk, or magnetic tape, the amount of data is enormous. Therefore, in order to record many frame images in one area of limited storage capacity, it is necessary to perform some kind of highly efficient compression on the data of the obtained image signal. Furthermore, in digital electronic still cameras, captured images are stored as digital data on storage media such as memory cards, magnetic disks, or magnetic tapes instead of silver halide film, so they are stored on memory cards, magnetic disks, etc. Alternatively, the number of images that can be recorded on one roll of magnetic tape device is specified, and the recording of this specified number of images must be guaranteed. Must be constant. Similarly, digital VTR (video tape recorder),
When recording moving images such as digital moving image files, it is necessary to be able to record various amounts of frames without being affected by the amount of image data per frame. In other words, whether it is a still image or a moving image, it is necessary to be able to reliably record the required number of frames, and the time required to record and reproduce data must be short and constant over a month. be. A coding method that combines orthogonal transform coding and variable length coding is widely known as a highly efficient compression method for image data. A typical example is the method being considered in the international standardization of still image coding. An outline of this method will be explained below. First, image data is divided into blocks of a predetermined size, and each divided block is subjected to orthogonal transformation. A two-dimensional DCT (discrete cosine transform) is performed. Next, linear quantization is performed according to each frequency component, and this quantized value is encoded as a variable length code.
1. Perform human encoding. At this time, regarding the DC component,

[1. in
Huffman encodes the difference value with the flow component of the neighboring block
do. The AC component is a low frequency called zigzag scan.
Scans from several components to high frequency components and disables (value
or 0) and the number of subsequent effective components.
Perform two-dimensional Nofman encoding from the values. With just this basic part, Huffman code, which is variable length encoding, is
Because encoding is used, the amount of code is not constant for each image.
turn into. Therefore, the following method was proposed as a method for controlling the amount of code.
ing. First, while processing the basic part mentioned above,
Determine the total amount of code generated for the surface. The DCT coefficient is calculated from this total code amount and the target code amount.
The optimal quantization width to get close to the desired code amount is determined.
Predict. Next, using this quantization width, the amount of the basic part is
Repeat the process after childization. 5.The total number of codes generated this time and the total number of codes generated last time.
The target code amount is again approximated from the code amount and the target code amount.
Predict the optimal quantization width for Then, this predicted quantization width and the previous quantization width are full.
The travelogue that occurred this time is closer to the minute and the amount of code is higher than the target code amount.
If the number is less than or equal to
Ru. If not, process using a new, higher quantization width.
repeat. Quantization width is the relative quantization characteristic for each frequency component.
Prepare a quantization matrix, which is the basic form representing
Required by multiplying the quantization matrix by the quantization coefficient
obtain a quantization width. Specifically, first the standard quantization
The quantum in the basic part mentioned above is determined by the quantization width obtained using the number
This is then variable-length encoded7, resulting in
information on the total code amount and the preset limit that should be stored.
The target total code amount is compared with the target total code amount.
When the quantization width is reached, the final encoding process is performed using that quantization width.
If the target total code amount is not within the target total code amount,
For example, linear prediction is performed from the total generated code amount and the target total code amount.
Find the optimal quantization coefficient that approaches the target total code amount by
, From the obtained quantization coefficient and quantization matrix,
Calculate a more optimized quantization width 7 and use this to create the final
Performs encoding processing. In this way - Toca method, the quantization width is
Make changes. However, with the code amount control method described above, the basic part of encoding
Because the number of times the minute bus is repeated varies depending on the image.
Not only is the processing time unstable, but the processing time is longer than that required for boat fishing.
There was one drawback that it required a long period of time. Therefore, the inventors proposed the following method to solve this problem.
proposed a method like this. (Special application flat), -2837B
No. 1, see Japanese Patent Application No. 137222/1999)
The compression method is a combination of orthogonal transformation and variable length encoding.
In order to control the amount of code generated in the formula, image memo is
Divide the sampled image signal stored in the memory into blocks.
, After performing orthogonal transformation for each divided block,
After quantizing this conversion output with a provisional quantization width, this
In addition to variable-length encoding the quantized output,
Calculate the generated code amount and the total generated code amount for the entire image, and then
The provisional quantization width, the total generated code amount, and the purpose.
A new quantization width is predicted from the total code amount. Also
, the amount of codes generated for each block, and the total number of codes
The allocation for each block is calculated from the amount of code and the total code amount
Calculate the amount of code. Then, using a new, higher quantization width,
Then divide the image signal in the image memory into blocks again and perform orthogonal transformation.
, quantization, and variable length encoding, as well as the generation of each block.
If the raw code amount exceeds the allocated code amount for each block,
Stop variable length encoding midway and start processing the next block.
Move. This allows you to determine whether the total amount of code generated for the entire image or the total amount of code for the purpose is
If you do not exceed the code amount, first, as shown in (a),
One frame of image data (as exemplified in the international standardization proposal)
The frame image is 720 x 576 pixels)
A block of a predetermined size (for example, 8x8 pixels or more)
) into blocks A, B, C, etc., as shown in (b).
As an orthogonal transformation for each divided block,
Performs two-dimensional DCT (fi scattered cocosine transform, 8×8
are stored sequentially on a matrix. Image data is two-dimensional
When viewed from the back side, frequency information based on the distribution of gradation information can be seen.
It has a spatial frequency that is a signal. Therefore, by performing -F DCT, the image data becomes
It is converted into a direct current component DC and an alternating current component AC.
There is a DC component DC at the origin position (0,0 position) on the
There is a detector indicating the value of , and at the 0.7 position there is a
The data indicating the maximum frequency value of the alternating current component AC is, and
, the maximum frequency of the alternating current component AC in the vertical axis direction is at the 7.0 position.
Furthermore, the data indicating the value is diagonally crossed at the 7.7 position.
Data indicating the maximum frequency value of flow component AC is stored respectively.
At the intermediate position, the relationship (1
Frequency data in the direction of eclipse is displayed sequentially from the origin side.
It will be stored in such a way that higher frequencies will appear.
. Next, the stored data for each coordinate position in this matrix
By dividing by the quantization width for each frequency component,
Linear quantization is performed according to each frequency component (C), and this amount
Variable length encoding is applied to the child value (7) and Huffman code is applied.
make a change. At this time, regarding the DC component DC, the neighboring block
The group number (number of additional bits) is
) and additional bits, and the group number is expressed as
The resulting codeword and additional bits are combined and encoded.
as encoded data (di, d2. el, e2)
. It is also not valid for alternating current component AC (value is not “0”).
) Coefficients are expressed by group numbers and additional bits. Therefore, the alternating current component AC is called zigzag scan.
Scans from low frequency components to high frequency components.
, the number of consecutive invalid (value 0') components (the number of zero runs)
), followed by the group number of the valid component value.
Two-dimensional Huffman encoding is performed, and the resulting code word and addition
The bits are combined to form encoded data. Huffman encoding is the above DC component D per frame image.
Occurrence in each data distribution of C and alternating current component AC
Focusing on the peak frequency, the more
, reduce the number of data bits to around 1-1.
Data is allocated in such a way that the number of bits is increased.
This is done by encoding the data [7] to obtain a code word. The above is the basic part of this method. In this basic part, Huffman code, which is variable length encoding, is used.
Because encoding is used, the code is not constant for each image.
Therefore, we will use a method of controlling the amount of code, and for example,
, process as follows. First, using the provisional quantization coefficients, the predetermined quantization
Each cycle obtained by multiplying the matrix and the quantization coefficient
It is the same as processing the basic part above using the quantization width for each wavenumber component.
At the same time, the total code amount (total bits) and number of bits generated for the entire screen is calculated.
Melt (g). This total code amount, the target code amount, and the usage
An estimate of the DCT coefficients from the provisional quantization coefficients etc.
Predict the optimal quantization coefficient to get close to the target code amount
Sur (h). Next, using this quantization coefficient, (i)
Repeat the process after quantization of the basic part described above. and,
The total amount of code generated this time, the total amount of code generated last time, and the target
The target code amount, the quantization coefficient used this time, and the
From the quantization coefficient obtained, the target code amount is again subtracted by 1.
Newton-Raphson Itare to find the optimal quantization coefficient for
- Jizun (Nekeiton Raphson 11erra
tion). And this predicted quantization coefficient and the previous quantization coefficient?
Close enough and the amount of code that occurred this time (also
If the total amount of code is small, the process is terminated and the current occurrence is
Output the encoded data and store it on the memory card.
(f). If not, change the quantization coefficient to 1 and
The process is repeated using the new 7 quantization coefficients. However, with the code amount control method described above, the basic part of encoding
Because the number of times the minute bus is repeated varies depending on the image.
Not only is the processing time uncertain, but the processing time is also shortened to boat fishing.
The disadvantage was that it required a long period of time. Therefore, the present inventors
I solved this using the following method. First, using the provisional quantization coefficients, the predetermined quantization
The matrix is multiplied by the jl 'T-ization coefficient, and the result is
The basic part is processed using the quantization width for each frequency component.
At the same time, the total code amount (total number of bits) generated for the entire screen is calculated.
) and calculate the generated code amount for each block (gl). this
The total code amount, the target code amount, and the tentative code size used.
Approach the target code amount for DCT coefficients from conversion coefficients, etc.
Predict the optimal quantization coefficient for (hl). Also, total sign
code amount, target code amount, and generated code for each block
Calculate the allocated code amount for each block by proportional allocation from the amount.
(h2). Next, using the predicted quantization coefficients H)
, repeat the basic processing described above. That is, the image again
Divide memory image signals into blocks, orthogonally transform, quantize,
In addition to variable length encoding, the amount of code generated for each block is
If the amount of code allocated to each block is exceeded, it can be changed midway.
Stop long encoding and move on to processing the next block. Then, finish the above process for all blocks with
After that, output the encoded data generated this time and store it in the memory card.
(f). [Problems to be solved by the invention] As mentioned above, for example, digital electronic still cameras
Natonioi uses one memory card or magnetic disk.
Images that can be recorded on a disk device or a single magnetic tape, etc.
The number of images must be guaranteed, so
The image data is compressed and recorded, but this processing is not recommended for ease of use.
The processing time must be as short as possible and constant.
Ru. It is also desirable to be able to compress image data with high efficiency.
will be included. These are not limited to digital electronic still cameras.
This is also required in many other applications.
It is an item. Therefore, the above-mentioned compression method satisfies such requirements.
There is also the International Standards Laboratory 7j method, and in this method, the above-mentioned standard
Orthogonal transformation and variable length for each block as illustrated in this section
Methods that combine encoding can highly compress image data.
Although it can be done efficiently, because variable length encoding is used,
The amount of code is not constant, and one memory card or magnetic
recorded on a recording medium such as a disk or a piece of magnetic tape.
The drawback is that the number of images that can be stored is indefinite. Ta. In addition, as a conventional technique that does not solve the two problems, the above-mentioned
Codes being considered for international standardization of still + L image coding
Is there an amount control method? In this method, the basic part of encoding is
The number of times the bus is repeated varies depending on the image, so
Since the fishing time becomes undefined, I decided to go fishing on a boat.
The drawback was that it took a long time. The inventors have proposed a method to solve this problem.
Japanese Patent Application No. 1-2837B1 and Japanese Patent Application No. 2-1.372
There is method number 22. This proposed method is usually
, very good results can be obtained or the system to be applied
Therefore, if the spatial frequency is particularly low ts or particularly high
The quality of the image deteriorated slightly7, and the amount of generated code was below the target level.
? The gap to the No. 1 amount is slightly larger than usual.
This happens very rarely. The cause of this will be explained. As mentioned above, the amount of change is 1.
is the basic expression that represents the relative quantization characteristics for each frequency component.
The quantization matrix in the form 1 is multiplied by the quantization coefficient.
When expressed in terms of
There is another relationship. log BR-a log SF+ b-(1) where
, BR is the code per pixel Jfl (bi-sotorate)
, SF is a quantization coefficient, a and b are constants, and a is a special
It is highly dependent on the system, and the image quality is very high.
Depends on it. When quantizing a color image separated into color components, the color components
For example, multiple quantization matrices, e.g.
and two quantization matrices for color difference and
, if a common quantization coefficient is used, then 10g BR-a log sp+ b ,
-=(2)y y 1ag BR-a log SF+ b c
... 3) The relationship will be like CC. In addition, in each of the above formulas, BR. Add y5C to a and b to distinguish between brightness and color difference.
It is. Here, if a − a, then ! og(BR1BR) c by bc -a log SF+(10+IO)-a log
Like SF+ b - (4)
The amount of code allocated to each block is determined using the statistics.
Is it okay to obtain the target code amount by proportionally distributing it according to the
Recognize. In the normal case, aZ:a or the system y
Depending on c, in very rare cases the size of a and a may be quite large y
c There is something different. In other words, the total code amount log BR is plotted on the Y axis, and the quantization coefficient l
og SF on the X-axis and the relationship between equations (2) and (3)
The result is as shown in Figure 10, and the quantization coefficient log 5P
The magnitude of the total code JllogBR according to IN= is the luminance system and
There is a slight difference in color difference. If you look at Figure 10, you will see that
The quantization coefficient when encoding, that is, the predicted optimal value.
If the quantization coefficient is particularly large or small,
In other words, especially for images with high spatial frequencies,
In the case of low or low images, the statistics of each block
Allocate by proportionally distributing the target code amount according to
Once the code amount is determined, the predicted quantization coefficient is used to
When actually encoding the luminance component block,
If the number of blocks of color difference components is
Then it becomes redundant, or vice versa. Is there a shortage of allocated code for the luminance component or for the chrominance component?
Whether there is a shortage of allocated code amount in a:2a
Size matters! :, % r - measured m f coefficient V
C Depends on the total size. , From the cause, to the applied scenario,...
For images with particularly low or particularly high spatial frequencies, young
The image quality deteriorates and the amount of code generated is less than the target amount of code.
In some cases, the gap was slightly larger than usual.
. Therefore, the purpose of the present invention is to obtain a spatial frequency distribution.
Even if the image has a skewed tendency, it can be used without sacrificing image quality.
Encoded within a certain amount of code within a certain processing time
Image data encoding device and encoding capable of
The purpose is to provide a method. [Means for Solving the Problem] In order to achieve the above object, the present invention is configured as follows.
. In other words, color image data separated by color component.
Divide each into blocks of a predetermined number of pixels, and divide each color component into blocks.
Perform orthogonal transformation for each divided block in order to generate frequency.
Convert this converted frequency component into separate data.
The data is quantized by quantization means, and then this
The quantized output of can be given to l'lJ variable length coding stage J'.
When encoding image data using variable length encoding, first
Temporarily quantize with the specified quantization width and encode with variable length code.
, perform statistical processing to obtain the code amount for each block, and then perform the following steps.
The code amount for each block obtained by statistical processing and the image
This total code amount target value is calculated from the total code amount target value of
optimal quantization width and amount of code allocated for each block
and each block is determined according to the determined optimal quantization width.
This quantization is performed while quantizing the conversion output for each block.
The output is divided into blocks according to the allocated code amount for each block.
Sequentially perform variable length encoding with i'=1 while controlling the code amount.
In the coding system that performs coding processing, firstly, based on the variable length coding output during the statistical processing,
Calculation means for calculating the code amount of each block for each color component
and each color component when the optimal sonization width is quantized L5.
For each "f-colorimetric colorimetric component code amount"
The total code; 1-] is distributed proportionally to the target value, and each color component is given a separate code.
Once the quantity is determined, this is used as the calculated value of the calculation means.
Therefore, the standard code of each block for each color component is proportionally distributed.
Determining means for determining the amount and encoding processing H4jζ
Determining means L 7 Determined reference code for each block for each color component
of the block is then encoded using the amount
In the processed block processed before the target block
The amount of code generated by encoding and the processed block
The difference between the allocated code amount and the allocated code amount is calculated and used as the
The code amount obtained by adding the code amount to the standard code amount of
is assigned as the allocated code amount for the target group τ-hook.
an allocation means for assigning and encoding processing in the target block;
the said allocation allocated by said allocation means in the process;
The above-mentioned variable-length encoding is discontinued so as not to exceed the code amount.
and terminating means. Second, the EiJ variable length encoding output during the statistical processing
Calculate the code amount of each block for each δ hull component based on
together with the total number of codes for each color component and the total number of codes for the entire image.
8 means, the total code amount target value and the % output
The optimum quantization width for each color component is determined by Y based on the calculated value of the means.
- a quantity to be measured J1 quantization width Y-II means and the optimum quantization width
Predict the code amount for each color component when quantized with 1, and calculate each of these
Compare the target total code amount based on the predicted code amount for each color component.
In addition to determining the amount of code for each color component,
Proportional distribution is made according to the calculated value of the calculation means, and each color component is
Determining means for determining the reference code amount of each block and encoding
Each block for each color component determined by the determining means during processing.
Using the reference code amount of the block, the next code amount of the block is
Processed blocks processed before the target block to be processed
Amount of code generated by lock encoding and the corresponding processing completed
Find the difference with the allocated code amount in the block and process it as described above.
obtained by adding it to the reference code amount of the target block.
The amount of code allocated to the block to be processed is the amount of code allocated to the target block.
the allocation means for allocating the
before being allocated by the allocation means in the encoding process.
Variable-length encoding is performed so as not to exceed the allocated code amount.
and terminating means for terminating. Thirdly, for each of the above first and second configurations,
Furthermore, select the desired compression ratio and
The target amount of code that should be accommodated according to the compression rate and the provisional amount
Compression rate selection means for determining the desired quantization width is provided.
. Also, fourthly, each of the above first and third configurations
Furthermore, the quantization means has a
Quantification is the conversion of each frequency component obtained by the orthogonal transformation.
When performed using data of low frequency components of data,
In both cases, the IjJ variable-length encoding means has the orthogonal transform.
The image data of each block obtained for each frequency component is
Of the data by frequency component of the motor, for the DC component,
In addition to obtaining the difference with the DC component in the previous processing block,
a DC component encoding processing unit that variable-length encodes the difference between
, after the encoding processing unit has finished processing this DC component.
, for the AC component to be variable-length coded for the AC component.
an encoding processing section, and the encoding aborting means is an AC
It is provided only in the component encoding processing section. Fifth, color image data separated by color component.
The data is divided into blocks of a predetermined number of pixels, and the divided blocks are
After sequentially performing orthogonal transformation for each block, the output of this transformation is
quantization means, and then this quantization output is variable.
The data is encoded by applying variable length encoding to the long encoding means.
When performing
and the percentage for each block determined from the statistics for each block.
The code amount is controlled for each block according to the code amount.
When performing variable length encoding, the quantization is temporarily
This is done using the quantization width, and the resulting quantization output is
Entire image required for encoding and optimization
The first step is to check the code amount for each block and the code amount for each block.
step and the code amount of the entire image obtained in this first step
The second step is to predict the quantization width required for optimization from
and each color when quantized with the predicted optimal quantization width.
The amount of code for each component is predicted, and the number of codes for each predicted color component is calculated.
Then, the total code amount target value is proportionally distributed5, and each color component is divided into
Determine the amount of code and apply it to the code of each block.
Proportional allocation based on quantity, standard allocation code for each block
The first step of calculating the amount and the predicted quantization width
a fourth step that performs the quantization for each block using
step and for each block in this fourth step.
The quantized output fits within the allocated code amount for that block.
a fifth step of variable length encoding in the range;
The block currently trying to perform variable length encoding processing in step
Variable-length encoding of the lock's most recently processed block.
Find the difference between the amount of code and the amount of code allocated for this block.
, use this as the carryover amount for the block that is currently being processed.
Add it to the standard allocated code amount in the lock, and add it to the current processing
The sixth step is to set the allocated code amount in the block to be
It consists of a top and a top. Sixthly, in the fifth case, the first step
In the step, the quantization is performed using a provisional quantization width.
, perform encoding processing on the quantized output obtained by this.
, the amount of code for the entire image and each color component required for optimization
Check the amount of code for each block and the amount of code for each color component.
In the second step, this first step
Optimization is performed from the obtained code amount for the entire image and the code amount for each color component.
In addition to performing f-measurement of the necessary quantization width for each color component,
In step 3, quantization 1 is performed using the predicted optimal quantization width.
7, predict the code amount for each color component, and calculate the amount of code for each color component.
proportionally allocating the total code amount target value based on another code amount,
Determine the amount of code for each color component, and apply this to each color component.
Proportional distribution is performed based on the code amount of each block, and each color component is
The standard allocated code amount for each block is determined, and the fourth step
In the step, the predicted quantization width for each color component is used.
perform the quantization for each block for each color component, and
In step 5, each color formation in this fourth step is performed.
The quantized output for each block is
In addition to performing variable length encoding within the allocated code amount,
In step 6, the current variable in the fifth step is
Processed immediately before the block to be long encoded
Code amount by 61 variable length coding in block and this block
Calculate the difference between the amount of code allocated in the
...The above in the block that is currently being processed
The block to be currently processed is added to the standard allocated code amount.
The procedure is to allocate the amount of code in the lock. Seventhly, color image data separated by color component
is divided into blocks of a predetermined number of pixels, and this divided block is
Perform orthogonal transformation for each block in order to create data for each frequency component.
At the same time, the frequency component obtained by this orthogonal transformation is
Separate data is sequentially divided into low frequency components by quantization means.
quantize and then apply this quantized output to variable length encoding means.
When encoding data by variable length encoding,
Total code amount target value to be stored and each block determined in advance
According to the allocated code amount for each block determined from the statistics for each block.
variable length code while controlling the code amount for each block.
When performing quantization, the quantization is performed using a provisional quantization width.
The quantized output obtained by this is encoded using
The amount of code for the entire image required for processing and optimization.
and the amount of code for each color component and the code amount for each block for each color component.
The first step is to measure the amount, and in this first step
Optimization is performed from the obtained code amount for the entire image and the code amount for each color component.
The second step is to predict the required quantization width for each color component.
and each color when quantized with the predicted optimal quantization width.
The amount of code for each component is predicted, and based on this amount of code for each predicted color component,
Then, the total code amount target value is distributed proportionally, and each color component is divided into separate codes.
In addition to determining the code amount, this is also determined as the code amount of each block.
Standard allocation for each block by allocating according to
A third step of determining the code amount, and the aforementioned predetermined Mj
Using the quantization width for each color component, each block for each color component is
a fourth step of performing the quantization for each block;
The quantized output for each block in the step is
Variable-length encoding is performed within the allocated code amount for the block.
each of the above obtained by the orthogonal transformation.
Regarding the DC component of the block-based image data,
Obtain the difference from the DC component in the previous processing block, and calculate this difference
variable-length encoding of the DC component, and encoding processing is performed for this DC component.
After that, the AC component is subjected to variable length encoding processing.
a fifth step, and the current state in the fifth step.
Immediately before the block to be processed with iq variable length encoding (7)
The code amount by variable length coding in the block processed in
Calculate the difference between the amount of code allocated in the block and carry it forward.
and the amount of the above in the block that is currently being processed.
Add the block to the standard allocated code amount and perform the current processing.
a sixth step of assigning the code amount in the lock;
The allocation in the fifth step? '1 quantity
The variable length encoding process in the range of
It is characterized by the fact that it can be applied. Furthermore, in the eighth method, in the fifth to seventh methods,
Note: Tentative quantization width of 1, total per selected screen
It is characterized by using a predetermined value according to the amount of code.
. [Function] In such a configuration, colors separated by color component
Divide each image data into blocks of a predetermined number of pixels,
Perform orthogonal transformation for each divided block for each color component.
Convert to data for each frequency component, and use this conversion to
The data for each frequency component is quantized by a quantization means.
, then this quantized output is given to variable length encoding means.
When data is encoded using variable length encoding, a temporary
Quantize with a fixed quantization width and perform variable length encoding2.
Statistical processing is performed to obtain the amount of code for each block, and then this
The amount of code for each block obtained through statistical processing and the image
The total code amount target value is within this total code amount target value.
The optimal quantization width and allocated code amount for each block
is determined, and each block is
quantizes the conversion output for each
power for each block according to the allocated code amount for each block.
Coding process that sequentially performs variable length encoding while controlling the amount of code.
However, in the case of the first configuration, the calculation means is
Based on the variable-length encoded output during measurement processing, each color component is
The code amount of the block is calculated, and the determining means is the optimal quantization.
The code amount for each color component when quantized by width is predicted7, and this
The total code amount target value is determined based on the code amount for each predicted color component.
Proportional allocation is made to determine the amount of code for each color component.
is distributed proportionally according to the calculated value of the calculation means, and each color is
Determine the reference code amount for each component block. Also, the quota
The means is configured to send each of the determined values to the determining means during the encoding process.
Using the standard code amount of each block for each color component,
Of these, the processing block is processed before the next block to be encoded.
code generated by encoding in the processed block
Find the difference between the amount of code and the amount of code allocated in the processed block.
Add this to the allocated code amount of the block to be processed.
Yes, the amount of code allocated in the block to be processed is
Teru. Then, the aborting means is the block to be processed.
The allocation allocated by the allocation means during the encoding process of
The variable length encoding is terminated so as not to exceed the specified code amount.
Ru. In addition, in the case of the second configuration, the p, page appearance, and column are the syntax
Based on the variable length encoded output during processing, each color component is
Calculate the lock code amount and total code amount for each color component
and calculate the total code amount of the entire image, and also calculate the quantization width
#j means is the total code amount target value and the calculated value of the calculation means
Based on the optimal amount J″′-′- conversion width for each color component1,
The determining means determines each color when quantized with the appropriate quantization width.
What is the code amount for each component? −The amount of code for each predicted color component
The total code amount target value is calculated based on each color component.
Proportional allocation according to another total code amount, and this is calculated according to the above calculation.
The code amount is distributed according to the code amount of each block for each color component.
By doing so, the standard code amount for each block for each color component is determined.
decide. In addition, the allocation means is configured to use the determining method at the time of encoding processing.
Using the standard code amount for each block for each color component determined in step
Among the blocks, the next processing target to be encoded
For encoding on processed blocks processed before the block
The amount of code generated and the assigned code in the processed block
Find the difference between the
The code amount obtained by adding it to the standard code amount is
Assigned as the allocated code amount for the block to be processed. Zoshi
, and the termination means is the encoding in the block to be processed.
The above assigned by the above two assigning means in the process.
In order not to exceed the allocated code amount, the i3J variable length
Controls tearing. [I-1-I Town, Tokyo 2] Using a provisional quantization width during processing
The total generated code L of the entire screen based on the quantization of
Generated code amount for each color component, code for each color component 1 for each block
Source 4 - Know the code amount status, and calculate the maximum J from this -1 information.
Measure the quantization width by t7, and quantize with the predicted optimal quantization width.
API L predicts the print amount for each color component when
The target value of the total code ff1tj is calculated based on the code amount for each color component.
Proportional allocation is made and the code for each color component is determined.
Distributed by the amount of 71 generated in each block of giJ for each color component.
to determine the standard allocated code amount for each block for each color component.
, when encoding each block in the encoding process,
The standard allocation code amount for that block [knock]
Difference between the code amount and the allocated n column when unifying the block
is added as the carryover amount and the assignment mark is 'QQ', which is 1.
It is said that it is encoded by enclosing i that fits in the allocation n number 2.
Based on data from unified 1 processing, each color component is
The standard allocated code amount for each block of
The distribution E5 is determined rationally according to the amount of code, and
The unused portion of the allocated code amount in the previously processed block is repeated.
5 to be used for encoding later blocks.
This is a color component that requires a large amount of code.
However, the number of coding truncation in each block is reduced, and the image quality is improved.
In addition to contributing to quality improvement, the number of encoding processes is reduced to a specified number of times.
The time required for encoding can be kept constant if
Ru. In addition, in the case of the third configuration, the above first and second cases
In addition, a compression ratio selection means is provided,
When you select the desired compression ratio, the compression ratio selection method
The desired amount of code to be stored according to the selected compression rate and
and determine the tentative fl T conversion width. Therefore, in this method, the functions and effects of the first and second configurations are
In addition to this, you can select the desired compression ratio.
In response to this, the provisional quantity should be adjusted to the width per image.
The desired code amount is determined and statistical processing is performed.
To more accurately find the optimal quantization width that fits within the
This gives you the advantage of being able to do more. In addition, in the case of the fourth configuration, the first and third configurations
In each case, the function of the variable length encoding means is
The image data of each block obtained by the orthogonal transformation is
Of the data, the DC component is processed in the previous processing block.
Obtain the difference with the DC component and transform and encode this difference
However, the encoding processing unit has finished processing this DC component.
After that, the AC component is variable length encoded and
Coding truncation is used in the encoding process for AC components.
We are using a configuration that is designed to understand and implement the system. Therefore,
The quantization means performs quantization at the time of the encoding process.
Among the data for each frequency component obtained by orthogonal transformation,
It is carried out based on data of low frequency components, and the variable
The long encoding means obtains each frequency component by the orthogonal transformation.
Frequency components of each block of image data
The DC component of the data is processed in the previous processing block.
Obtain the difference with the DC component of
, after the encoding process is completed for this DC component, the above-mentioned
Variable-length encoding is performed on the AC component, and the encoding
The truncation applies only to the AC component. As a result, in the first to third cases above, further encoding is performed.
The encoding truncation during processing was made only by illusion for the alternating current component.
This allows for the truncation of high frequency regions that have little visual impact.
It is possible to prevent deterioration of image quality. In the fifth case, color image data separated by color component is used.
Divide the data into blocks of a predetermined number of pixels, and
After sequentially performing orthogonal transformation for each lock, the output of this transformation is quantified.
It is quantized by the childization means 2, and then this quantized output is
Variable length given to variable length encoding means? : f code 5, data
The total amount of code that should be stored in advance when encoding data.
Each block determined from the target value and statistics for each block.
Code amount is controlled for each block according to the assigned code for each block.
When performing variable length encoding while performing
・Use the knob to perform the quantization using the provisional quantization width.
, perform encoding processing on the quantized output obtained by this.
, the amount of code for the entire image and each block required for optimization.
In the second step, the amount of code for each block is checked.
The quantum required for optimization is calculated from the code amount of the entire image obtained in step
In the third step, the prediction width is f-measured.
Code amount for each color component when quantized using the measured optimal quantization width
is predicted, and the total amount is calculated based on the code amount for each predicted color component.
The code amount target value is distributed proportionally to determine the code amount for each color component.
At the same time, this is compared based on the code amount of each block.
example, calculate the standard allocated code amount for each block, and
In the step, the predicted quantization width is used to
The quantization is performed for each block. And the fifth step
Each block in this fourth step
The quantized output for each time is assigned to the code amount for that pro-nk.
Either use variable length encoding to fit within the range, or use
The allocated code amount for the block currently being processed is
Immediately before the block to be currently variable-length encoded
The amount of code generated by variable length coding for blocks processed in
Find the difference between the allocation and the amount in the block and carry this forward.
1, the amount in the block currently being processed is
The amount added to the standard allocated code amount is applied. In this way, in the fifth case, we first set a provisional quantization width.
The total generated code amount for the entire screen based on the quantization using
and the generated code amount for each color component, and each block for each color component.
Knowing the state of the generated code amount, and using this information, the optimal quantum
When predicting the quantization width and quantizing with the predicted optimal quantization width
The code amount for each color component is predicted, and the code amount for each predicted color component is
The total code amount target value is distributed proportionally based on the amount of code, and each color component is
In addition to determining the amount of code for classification, this is applied to each block for each color component.
The code is distributed proportionally to each block according to the amount of code generated in the block.
Determine the standard allocated code amount for each block for each color component, and
When encoding each block in the encoding process,
The code of the previous block is added to the lock standard allocated code amount.
The difference between the code in
Add 1 to the allocated code amount 2, and fit into this allocated code amount.
It is said that it is encoded in the range j, t, and the unified knee 1
Based on the data during processing, each color component is divided into blocks
(The standard assigned vs. amount) reasonably according to the generated sign vs. amount
It is distributed and determined, and the block processed earlier is
The unused portion of the allocated code amount is carried over as 7 minutes and used as 5 for later blocks.
Since the code can be used to encode the
Even for color components that require a large amount, the sign in each block is
This reduces decoding truncation, contributes to improved image quality, and
, if the number of encoding processes is set to a specified number of times, the number of times required for encoding is
The period of time can be fixed. In a sixth case, in the fifth case, the first step
Using the provisional quantization width of 1j in step
and encode the resulting quantized output.
The amount of code required for the entire image and each color for optimization.
The amount of code for each component and the amount of code for each block for each color component.
The second step is to investigate this first step.
The code amount of the entire image obtained by
Perform lf-measurement of the quantization width required for optimization for each color component 2
- In the third step, the optimal quantization width is used for quantization.
Then, the initial amount of each color component is preliminarily iip + L, and each of these
Based on the predicted code amount for each color component, set the target value for the amount of confusion described above.
Proportional allocation is made to determine the amount of code for each color component.
is proportionally distributed based on the code amount of each block for each color component.
and calculate the standard allocated code amount for each block for each color component.
, in the fourth step, the predicted amount of each color component is calculated.
The quantum size for each block for each color component is calculated using the child width.
and in the fifth step, in this fourth step
The quantized output for each block for each color component of
Divide to the block = variable length within the range of i sign to quantity
At the same time, in the sixth step, the fifth step
The block that is currently being processed for variable length encoding in
, code by variable length encoding in the block processed immediately before
Find the difference between the amount of code and the amount of code allocated for this block, and calculate this as
In the block currently being processed as the carryover amount
is added to the standard allocated code amount, and the amount of code currently being processed is
The assigned code for the professional note shall be; In the sixth case, quantization is carried out to a provisional quantization width,
Encoding process for the fit “F ratio output obtained by this
The amount of code required for the entire image and each
Code amount for each color component and R'T for each block for each color component
ffl to key, and the code amount of the entire image obtained by this and each color
Calculates and measures the quantization width required for optimization from the amount of code for each component.
In addition to calculating each color component, we also calculate the total code that should be stored in advance.
Quantity target value, expected generated code amount for each color component, and each of the above color components
For each color component, each color component is allocated proportionally from the code amount of each block.
Determine the standard allocated code amount for each block, and then
The quantization width for each color component is used to calculate each block for each color component.
The above quantization is performed for each block, and here 11) M”J 2
The quantized output for each block is divided into that block.
Variable-length encoding is performed within the code amount. - force, of the block to be currently variable-length encoded;
Code amount due to variable length encoding in the block processed immediately before
Find the difference between
The current processing (,
In addition to the standard allocation fJ in the
2. Assuming that the allocated code amount is
To encode within this code amount range, use variable length encoding.
. 2, quantization using a provisional quantization width for the first time
The total generated code amount for the entire screen based on
Amount of generated code, amount of generated code for each block for each color component
Know the state and use this information to find the optimal quantization width using the f-measure.
In addition, the travel log amount 11 target value is adjusted according to the expected generated code amount.
5, and then distribute this to each color component.
Each block for each color component is distributed proportionally according to the raw code amount.
Determine the standard allocated code amount for each block, and
When encoding a lock, the standard assignment mark for that block is used.
The code amount in the encoding of the previous block is added to the code amount in the previous block.
The difference between the allocated code amount and the carryover amount [7 is added and 5 is the allocated code]
code amount, and encode L2 within the allocated code amount.
What is said to be - C1 processing using provisional quantization width
data at the time of calculation and the predicted code amount at the predicted optimal quantization width.
Based on C1 Standard assigned code for each block for each color component
The amount is rationally allocated and determined, and the amount is
The unused portion of the allocated code amount in the allocated block is carried over.
What can be done to encode later blocks?
Therefore, even for color components that require a large amount of code, each block
This reduces the number of coding aborts during storage, contributing to improved image quality.
At the same time, if the number of encoding processes is set to a specified number, the sign
The time required for encoding can be fixed. In addition, in the seventh case, the colors separated by color component are
Divide the image data into blocks with a predetermined pixel size, and
Perform orthogonal transformation for each divided block to separate frequency components
In addition to converting the data into
The data for each frequency component is converted into low frequency components using quantization means.
Quantize sequentially from minute to minute, and then convert this quantity j°−
Variable-length encoding is performed using variable-length encoding means to encode the data.
The target value of the total code amount that should be stored in advance when performing encoding.
Determine from the statistics for each block 1 and each block
Control the code amount for each block according to the allocated code amount for each block.
In performing variable improvement while performing, the above-mentioned Hr improvement
Using the provisional quantization width, the obtained quantity f
Perform encoding processing on the specific output and use the necessary
The amount of code for the entire image, the amount of code for each color component, and the amount of code for each color component.
The image obtained by examining the code amount for each block of classification
Amount required for optimization based on overall code amount and code amount for each color component
Prediction of the child width is performed for each color component, and then the optimal quantization is performed.
The code amount for each color component when quantized by width is predicted, and each
Compare the target total code amount based on the predicted code amount for each color component.
As well as determining the amount of code for each color component,
Proportional allocation based on the code amount of each block for each color component
Find the standard allocated code amount for each block for each color component.
. Then, the sonization width for each color component predicted by the previous λ is
perform the quantization for each block for each color component using
At that time, the quantized output for each block is
Variable-length encoding is performed within the allocated code amount for the block.
Ru. At this time, each block obtained by the orthogonal transformation
Of the image data for each block, the DC component is
Obtain the difference from the DC component in the processing block and use this difference as
Variable length encoding 5, and °C encoding processing for this DC component
After -9, +1]' variable length for the AC component
Encoding process and current variable length encoding process in the second process
In the block processed immediately before the block to be
The code amount by variable length encoding and the assigned code in this block
Calculate the difference between the current amount and the carryover amount.
In addition to the standard allocated code amount for the block to be processed.
and the assignment mark in the block that is currently being processed.
The quantity shall be the number. Then, within the range of the allocated code amount,
The variable length n9 processing is applied slowly to the AC component.
. In this way, the first cut in variable length encoding is
component only, and the low frequency component that has a large visual impact
Since truncation is suppressed, image quality can be maintained. Furthermore, eighthly, in the cases 5th to 7th above,
, the provisional quantization width is set per selected screen.
A predetermined value of 7°ζ is used depending on the amount of travel. As a result, the provisional sonization width is the same as the width of the screen that should be accommodated.
It is close to the suitable value for frogs depending on the amount of travel, and is more appropriate.
Compression encoding becomes possible. As shown in 2, the present invention first performs four treatments on eel, and this process ≧
-1 processing predicts the optimal quantization width for each color component,
Necessary for determining the standard assignment code I for each block for each color component
Obtain the statistics and start the encoding process based on this information.
, the standard code amount for each block considering the predicted quantization width.
Process while allocating the
The standard allocation code amount of the block to be used and the block's
In addition to the excessive amount of code allocated in the previous processing block,
Within the allocated code amount of the block to be processed
block by discontinuing variable length encoding so that the
The code amount is controlled for each code, and the code is within the desired code amount.
The converted output is obtained as the final output. This is based on the following idea. Explanation using a conventional example (7) l1lr when encoding
When the r conversion coefficient is particularly small or small, paraphrase
and especially for images with high or low spatial frequencies.
Then, we compare the code amount of Li according to the statistical amount of the name 70.
Example 4: When determining the allocated code amount by
Actual encoding is performed using the Y-measured QF coefficients.
When this happens, the allocated code amount for the luminance component block is
Insufficient and redundant in the chrominance component block, or
The opposite may be true. Then, the amount of code is insufficient.
Of course, the encoding of the color component is aborted midway through.
Image quality deteriorates. Whether this phenomenon occurs or not, the color
Whether the amount of allocated code for the luminance component is insufficient or not
, Is the allocated code amount insufficient for all color difference components in the color components?
Whether it is fortune-telling or not depends on the magnitude relationship between a in equation (2) and a in equation (3).
It can be easily determined by the magnitude of the predicted quantization coefficient.
Is this possible? This means that it depends on the system and the image.
It is determined. Therefore, the quantization width for each color component was predicted from the above relationship.
The code amount for each color component when actually encoded using
For each color component of each standard assignment code for each block;
The ratio is the same.
When (T), the emission for each color component/1'R month m is J'', 2
Equation (2,), Equation (3) Prediction of r-#IL, , :
The ratio of the generated code amount for each color component is the target travel amount.
After allocating, statistics for each block in each color component.
Further proration based on quantity. In this way,
In addition to determining the standard allocated code amount for each block,
The remaining amount of code allocated in the block that has been encoded is calculated as follows.
Follow this by passing the assignment code I of the next block.
Encoding is performed while controlling the amount of code for each block.
give The code is actually encoded using this predicted quantization width for each color component.
For each block, the ratio of the amount of code for each color component when
By adjusting the ratio of the standard allocated code amount for each color component,
The amount of code allocated to each block is determined using the predicted quantization width.
Amount of code generated for each block when actually encoded
The result is very close to
Ru. The prediction of the quantization width for each color component is based on equations (2) and (3).
Just do linear prediction. Therefore, according to the present invention, images with a wide range of spatial frequencies can be obtained.
On the other hand, if you want to reduce the image quality as much as possible, use a simple circuit.
Sign so that it falls within a certain η logarithm within a certain processing time.
become able to become [Examples] Examples of the present invention will be described below with reference to the drawings.
Ru. When compressing and encoding color image data,
Quantize with the predicted optimal quantization coefficient and encode it
For each color component such as luminance component and color difference component, the specified
If each is encoded to fit within the code amount range,
In this case, depending on the color component, there is a margin in the code amount, and on the other hand, the actual
In addition to fertilizing, it also grows, and in places like this,
In this case, high frequency components are cut on the side of the missing color component.
This happens one after another with one color component.
There may be problems such as the amount of code being too large rather than
Rationality arises. The present invention first uses color image data separated into color components.
We perform statistical processing on the
Predict the width of change, and use this statistical processing to find the optimal amount f-
Necessary for predicting the width and determining the amount of code allocated to the base ring for each block.
Obtains 1 amount of necessary engineering. And based on this information,
For each color component (for Y, for C or for Y, for Cr, for Cb)
) prediction of the optimal quantization width and the amount of code allocated to each block.
After making the decision, the actual encoding process is performed. At this time, the standard allocated code amount for each block is determined as described above.
Actual encoding using the predicted quantization width for each color component
When the ratio of the code amount for each color component is
Prediction is made so that the ratio of quasi-allocated code amount for each color component matches.
Occurrence code for each color component when encoding with the above quantization width
The amount of signal is predicted from the above equations (2) and (3), and this prediction
Allocate the desired amount of travel by the ratio of the amount of doto codes for each color component.
based on the statistics for each block for each color component.
This is done by further distributing the amount on a pro rata basis. In the encoding process
Then, quantize in block order using the quantization width for each color component.
And encoded block text (7) 111 hit
The remaining amount is used to allocate the amount of code for subsequent blocks.
, while making some adjustments to keep it within the allocated code amount.
While controlling the code amount for each block, the LiJ variable length n
Implement the code. First, in order to make the present invention easier to understand, we will explain the basics of the present invention.
Let me explain my thinking. That is, the present invention, for example,
One color image (screen) is divided into 14 color components, and these
Color image data for each color component is divided into blocks of a predetermined number of pixels.
Then, each block by color component is divided into blocks.
The data of the constituent pixels of that block are orthogonally transformed and
Separate the wavenumber range, further quantize it, and compress and encode it.
Then, the travel pair amount of each color component compressed and encoded is the target code amount.
In order to encode the color so that it fits within the
Statistical processing is performed on the separated color image data, and this
Through statistical processing, we can predict the optimal quantization width for each color component.
, the assigned code for each color component based on the generated code amount for each color component.
Proportional allocation of color and standard allocation mark for each block for each color component
Obtain the statistics needed to determine the quantity. And these circumstances
Based on the information, the standard allocated code amount for each block for each color component is determined.
The standard allocation R+3 amount for each block for each color component is determined.
code for each block for each color component so that it fits within
Perform conversion processing. The encoding process uses the predicted quantization width for each color component.
It is performed sequentially for each component and in block order, and the encoded output is
While looking at the standard allocation for each block? TJ; and its
The amount of code generated in the processing block “hook” before the processing block
Excess/deficiency amount for the allocated code amount (carryover amount)
within the allocated code amount of the block obtained by combining the
Monitor the amount of code obtained by variable length encoding so that
The amount of generated code, including %, is the allocated code.
When the amount of code is reached, the encoding of that block is terminated, or
The DC component must be encoded even if it exceeds the allocated code amount.
, stores the excess/deficiency amount for the allocated code amount at that time.
and then moves on to encoding the next block.
be. Statistical processing is used to predict the optimal quantization width for each color component.
In addition, the code amount for each color component is optimally allocated and further kneaded.
Determine the standard allocated code amount for each block for each color component based on
This is the only way to obtain the standard J1 quantity, which is the standard for determining the
Ru. The optimized quantization width obtained by the Toko 1 process is expressed as follows:
By using it in the encoding process, it becomes the general code for the purpose of
The upper limit of the amount of data for one image is regulated.
If it is set, not only 1.1 byte but also 1 bit
It is not possible to exceed the target code amount. Therefore, control is performed based on the allocated code amount for each block. child
The code amount at the time of encoding is the target code amount.
This is used for fine adjustment when changing the settings for each block.
The standard allocated code amount, which is the standard allocated code amount for each lock, is
This standard allocation code amount is used for the previous block.
Subtract the generated code amount from the allocated code amount in the processing block.
, the remainder after subtraction is received as a carryover amount, and this carryover (2
, the allocated code amount corrected by adding
This is the maximum limit for encryption processing, except in special cases.
Encoding control is performed to encode within the range. Adjustment of the allocated code amount due to this carryover is performed after encoding.
Since the amount of code differs for each block, the remaining
In the block where there is a shortage, the surplus is transferred to the block where there is a shortage.
Flexible and rational to fit within the target code amount as a whole
This is for the purpose of allocating the amount. This enables the standard code amount to be turned off.
This will give you the ability to make fine adjustments when changing the settings. each
From low frequency components to high frequencies in block encoding
The guideline (Cobro
When the amount of code allocated per block is exceeded, further
Aborts encoding of high frequency components. This is done for each block to minimize visual impact.
This is to omit the high frequency components first. This interaction
Regarding truncation of flow components, high frequency components are particularly
There is a tendency that the higher the wave number, the less the visual impact.
They are using the opposite. Amount of code generated including EOB or assigned code
The amount was less than the amount, or the amount was the same.
The code finishes encoding without any problems, that is, outputs EOB.
do. In this way, when there is a surplus, it is necessary to move on to the next block.
In addition to the lock standard allocated code amount, λ is allocated in this block.
Let it be the amount of code. Encoding is discontinued for blocks that overlap in the middle.
, finishes encoding that block, that is, outputs EOB
do. This is not possible in reality unless the compression ratio is extremely high.
In the unlikely event that the given allocated amount of code encodes even the DC component,
If the amount is not enough, encode only the DC component.
, finishes encoding that block, that is, outputs EOB
do. Then, the insufficient amount of code is treated as the missing amount.
block, and the shortfall is carried over to the base of that block.
Subtract it from the quasi-allocated code amount and use it as the allocated code amount for that block.
Ru. At this time, EOB is also one of the codes, so EOB
The amount of code including the amount of code should be within the allocated code amount. In this way, the allocated code amount is not wasted.
It is also an object of the present invention to reduce the amount of generated code to the allocated code amount.
If the surplus is carried over to the allocated code amount of subsequent blocks,
Both are the ratio of statistics for each block when statistics are obtained.
Predicts the quantization width for each color component so that the
By doing this, each block is
The amount of code allocated for each color is more reasonable, and each color
Good results can be obtained with both ingredients. The present invention basically consists of "statistical processing" and "encoding processing".
A method called the 2-bus method, in which encoding is completed by combining two
is adopted, and this is quantized with a certain quantization width.
When variable-length encoding is performed, the base assigned to each block is
Using the quasi-allocated code amount as a standard, the generated code in each block is
It is possible to absorb variations in the amount of data in the next block.
We are going to adjust this by assigning this adjusted value.
The code is I, and the code in the block is within this allocated code amount.
When the allocated code amount is exceeded, the block is
Encoding should be aborted. However, the image data is
It is decomposed into a DC component and an AC component by AC transformation, and the visual
The amount of code is tentatively assigned to the DC component that has a large impact on the
Even if the number of codes exceeds the number of bits, partitioning will be performed for the next block.
Subtract the amount of code allocated by the network by the amount that exceeds the allocated code amount,
In addition, there is a risk of entering the AC component and exceeding the allocated code amount.
- If the error occurs, the encoding is terminated. Also, by encoding
If the generated code 4-i amount is less than the allocated code amount,
The surplus is added to the allocated code amount of the next block, and
Increase the amount of code allocated to the block. In this way
, the code amount in the previous processing block or the allocated code amount is burned.
If the excess amount is returned to the next block,
This is reflected in the allocated code amount for the network and adjusted accordingly. In general, especially for images with high or low spatial frequencies
In, each block's lesson;
By allocating proportionately the amount of travel to the purpose to be accommodated.
Determine the standard allocated code amount for each block, and calculate each predicted color.
Actual encoding using one quantization coefficient common to all components
If the conventional method is adopted, the luminance component block
If the amount of allocated code is insufficient, and the block of color difference components
It may be redundant or vice versa;
For color components with insufficient signal amount, encoding is aborted midway.
If this happens, the image quality will naturally deteriorate.
According to
The main reason for this is that the internal circuit allows for
Encoding can be performed within a certain amount of code. An embodiment of this device using the above principle will be described. Figure 1 shows an image data encoding device according to the present invention.
An example applied to a digital still camera is also described below.
Figure 2 shows the configuration of the image data encoding device according to the present invention.
Each is shown in a block diagram. Note that this invention is not directly related to the present invention.
The mechanism of a digital electronic still camera without
The explanation will be omitted. As shown in Figure 1, the digital electronic still camera body
(hereinafter referred to as the electronic camera body) 1 captures images
A predetermined signal is applied to the imaging system 40 and the output of this imaging system 40.
A signal processing circuit 60 that performs signal processing, orthogonal transformation, linear quantum
It has functions such as conversion, fil variable length encoding, etc., and
An encoding circuit 8 outputs the output of the circuit 60 through compression encoding 1 to
0 and the image data encoded by this encoding circuit 80.
records data and quantization width (or corresponding information)
Recording system 70 for recording on medium 7I and desired data compression rate
Switch 30, which is manually set, controls the entire system.
It is composed of a control circuit 90. Set the image compression rate on the control panel of the electronic camera body.
A switch 30 is provided, and the switch 30 controls
It is connected to circuit 90. The imaging system 40 includes a lens 40a for forming an optical image.
and an image sensor 40b such as a CCD. The signal processing
The logic circuit 60 is an amplifier BO that performs amplification, noise removal, etc.
a and convert analog signals to digital signals ^/D
The converter 60b and )? Buffer memory f consisting of AM, etc.
ile and a process circuit GOd that performs color signal formation, etc.
Be prepared. The encoding circuit 80 is, for example, a DCT (discrete code).
Orthogonal transform circuit 4 that performs orthogonal transform such as sine transform), linear
Quantization circuit 6 that performs quantization, Huffma as variable length encoding
You can see the Huffman encoding circuit 8 that performs quantum encoding, and
Converting width - Y - measurement circuit 12° code amount calculation circuit 14, code amount division
In this circuit 20, code terminating circuit 1B and encoding circuit 80
It has a control circuit 18 that performs control processing. The recording system 70 includes an interface circuit 70a and a recording
Memory card with built-in IC memory used as a medium
71. Memory card 71 is the electronic camera body
1, it is removable. The control circuit 90 is controlled by a microprocessor (MPtl).
It has been realized. FIG. 6 shows a perspective view of the external appearance of the electronic camera body. The diagram is
A binocular type one is shown, and 48 indicates 1 in the operating section.
, CI) (liquid crystal) display, 30 is the screen on the operation section.
switch 30, and also a tele/wide switch.
A switch, a shutter operation button 50, etc. are provided. Ma
In addition, 49 is a finder. 1-5CI) The display 48 is under the control of the control circuit 90.
Various values such as shooting mode, number of frames, day 4 = j, time, etc.
and status will be displayed. With this electronic camera,
Press the switch 30 provided on the operation part of the body 1 by pressing a f' +
In particular, it is possible to set the image compression ratio to the desired value.
I can do that. That is, the control circuit 90 is provided with a standard
Multiple types of compression ratio information are set, and this or switch
The value of the number of shots set by the operation in step 30 is also
The capacity of the memory card (recording medium) installed in J
Determine the compression rate to be applied from the amount, and calculate the value of the H compression rate found.
and converted to the value of the number of images that can be recorded on a memory card.
It will now be displayed on the LCD display 48 of the operation unit.
ing. Then, when the user presses the switch 30,
Each time the control circuit 90i;e switch 30 is pressed, this
Change the values of While viewing the displayed change values, the user selects the desired value.
Each time you stop pressing switch 3o, the control circuit changes.
Route 90 corresponds to the instruction value for the number of images taken at that time.
It is now possible to set the indicated compression ratio. This is pressure
Controls the standard travel amount per image determined according to the reduction ratio
The circuit 90 obtains this and uses it as the objective n jlia constant information.
This is done by applying the same to the encoding circuit 8o. Also, trigger gas
The shutter operation button 50, which is the switch, is pressed.
As a result, the shutter function moves and the image sensor 40b is exposed to light.
An image of the subject is formed, and the image sensor 4()b corresponds to this image.
Then, the image is stored, so read it out and control it.
It is possible to obtain a video signal from the image sensor AOb.
. These controls are also controlled by the control circuit 90 or 5. The imaging system 40 in FIG. 1 includes a photographic lens 40a and a CC
1, having an image sensor 40b made of an image sensor such as D;
An image is formed on the image sensor 40b by the photographing lens 40a.
The obtained optical image is converted into an image signal and outputted to the signal processing circuit 60.
It is something that gives strength. The signal processing circuit 60 is an amplifier 60a
SA/D converter 60b, buffer memory 60C1 process
A process circuit 80d is included, and this process circuit Hcf
The image signal obtained by the image sensor 4Ωb is converted into a color signal.
No. Y, RY (Hereinafter, this 1? -Y is Cr(
(abbreviated as Chroma Red), B-Y (hereinafter referred to as 20B)
-Y is abbreviated as Cb (chroma blue)) each color component
In addition to separating the
It is recommended that you do the following: Imaging system digitally converted by A/D converter 60b
For example, the output video signal of 40 has a capacity for one frame.
Image data is stored in a buffer memory 60e having
By being read out and input into the process circuit 60d,
The V component of the luminance signal system and chroma (C5 color difference signal)
The Cr and Cb acid components of the thread are separated. buff 7 memo
60e 1.1: The stored image data is
First, in order to perform a unified process on the f-number of the luminance system, we
The Y component of the image signal is processed by the process circuit.
Y component data is obtained and given to the encoding circuit 80.
After the encoding process is completed,
Next, we will discuss the chroma-based Cr and Cb acid component data.
After processing, encoding processing is performed. The f5 processing circuit 60 has a blocking function and a buffer
The data read from the memory 60c is processed by the L1 process.
Obtained image data for Y component and Cr, CI) component
(17 reno, minutes, young, < is 1 field),
Performs blocking processing to divide into blocks of a predetermined size.
I can do it. Here is an example (7) and the block size is
The block size is 8 x 8 pixels, but the block size is limited to 8 x 8.
It is not a block sensor with Y and C (chroma system).
The size may be different. In this embodiment, the luminance system Y data is read and divided into blocks.
Then, the subsequent processing system uses ljλ to process this Y component data.
After the statistical processing is completed,
Next, statistics on the data of chroma-based Cr and Cb components.
Data on the chroma-based Crx and Cb components to be processed.
Enter reading and blocking. To create chroma blocks, first Cr component image data is
Perform all blocking for , then cb formation
Block image data? It shall be carried out. The encoding circuit 80 has the configuration shown in FIG. No.
In Figure 2, 4 is an orthogonal transform circuit, which is divided into blocks.
Receive each image data inputted by
2, perform two-dimensional orthogonal transformation for each block.
It is something. Let's assume orthogonal transformation, then cosine transformation, sine
Transforms, Fourier transforms, Hadamard transforms, etc. can be used.
. By performing orthogonal transformation, the image data as transform coefficients can be
data can be obtained. The image data as this conversion coefficient is directly
Since it is an alternating transform, the spatial frequency of the image can be converted into a frequency component.
The data is analyzed every minute. 6 is a quantization circuit, which outputs the orthogonal transform circuit 4.
When receiving the image data (transform coefficients), the quantum of the first L]
In , each frequency component is set in advance for each color component.
For each quantization width, the amount set in advance according to the shooting
The amount of conversion coefficient is the quantization width corrected by multiplying by the childization coefficient a.
Childization is performed, and the second time it is determined by the previous processing.
The configuration is such that quantization is performed using the optimal r-ization coefficient α, and
Ru. 8 is a variable length encoding circuit, and the variable length encoding circuit 8 is a variable length encoding circuit.
The quantized output output from the slave circuit 6 is variable-length coded.
It is something that Huffman coding is used as variable length coding,
Use arithmetic encoding. Block in variable length coding
The amount of code for each image, the amount of stone for the entire image, etc. change for each image.
. The type of variable length encoding to be used is directly related to the present invention.
Although unrelated, I used Huffman encoding here.
Let me give an example. The variable length encoding circuit 8 converts the input quantized transform coefficient into
Zigzag cedar with numbers in the order shown in Figure 7
By using a method called
Performs scanning to wavenumber components. Data of the first DC component N)C] in the scanning order in Figure 7.
is the DC component of the block that was previously variable-length encoded.
Output the difference value with 1-fuman encoding. AC component
For [AC], 64 from the second in the scanning order in Figure 7
Look at the conversion coefficients in order until the conversion coefficient is 0.
When a valid (i.e. valid) coefficient appears, it is
The number of consecutive 0 (invalid) coefficients that existed (zero run) and their
Two-dimensional Huffman encoding is performed using the effective coefficient values of
It does the following actions. In addition, invalid coefficients are continuously displayed after a certain coefficient up to the 64th coefficient.
If the number continues, EOB (Enter) indicates the end of the block.
Outputs the code of block (de of block). Also, discontinued
When a signal is input, encoding is finished and EOB is added.
and output it. And the code generated for each code
The length is output to the code amount calculation circuit 14. The code amount calculation circuit 14 inputs each color of Y, Cr5Cb.
Y, Cr, C
b Collection of code amount data for each block of each component and 8 packets
Calculate the amount of code for each component and the amount of code for the entire image, and
Regarding data on the amount of code for each color component and the amount of code for the entire image
In addition to outputting to the quantization width prediction circuit 12, the
For the data of the code amount and the code amount of the whole image, divide the amount by n.
The configuration is such that the output is output to the applying circuit 20. The quantization width prediction circuit 12 performs control at the start of the first bus.
Information on the target code amount is received from the circuit 18, and this code is
Log (BRy → BRe) ; a X l based on the equation (4) explained earlier from the quantity information
og SF+b However, BR is a bit per pixel with a code amount of 0.
SF is a quantization coefficient, and a and b are constants. Set the initial value of the quantization coefficient α using the relationship 11, quantum
output to the converter circuit 6, and the signal is output to the history prior to the start of the second node.
, n vs. calculation circuit 14 to manually calculate the sign of the entire image.
amount and the maximum allowable input per image, '7) data amount.
From a certain target edge code amount, for example, <2) formula, <3) formula
Using the relationship, approach the target edge code amount by linear pre-Mj
The optimal Q f coefficient a for each color component was actually calculated this time.
Use (7, ffi 'Although it is predicted by taking into account the F coefficient
It is. Here, how to set the quantization coefficient to an optimal value is
Since this is an important issue, I will explain this point a little. Preprocess the image data, quantize this output, and
When variable-length encoding the encoded output, the quantization width of this quantization
It is well known that changing the amount of code will change the amount of code generated.
It is being This is a variable method represented by Huffman encoding.
Long encoding uses the bias in the probability of occurrence of the data to be encoded.
to reduce the amount of code required to represent that data.
Therefore, the above-mentioned ``changing the quantization width''
"to change the probability of occurrence of the quantized value"
Therefore, by changing the quantization width, the
It can be seen that the raw code amount also changes. By the way, even if the same encoding is performed with the same quantization width, 1.
The amount of generated code varies depending on the image data at that time. L
The quantization width is changed for the image data of
If the same 6-encoding is performed, the quantization width and the generated code amount
A certain relationship can be obtained between Also, many image data
When determining the relationship between the quantization width and the amount of code emitted by the
It is clear that a relationship with a high degree of raw axis can be obtained statistically.
It was. Specifically, in many cases, the following relationships were obtained. In other words, let SF be the relative ratio to a certain quantization width.
, generated code amount, number of bits per pixel (bit rate)
If this is expressed as BR, then the relationship in equation (1) above is established.
stand up In equation (1), if a is the same encoding, then
, is approximately constant regardless of the image, and b depends on the image. child
The value of b has a certain distribution depending on the image, and this occurrence frequency
The typical value can be obtained from the degree distribution. Therefore, find the optimal quantization coefficient from the relationship in equation (1).
I can do it. The code amount allocation circuit 20 receives input from the code amount calculation circuit 14.
Code amount of image data for each color component, product for each color component
Code amount for each block, code amount for the entire image, and target travel number
Calculate the standard allocated code amount for each color component and product block from the amount.
The signal is then output to the encoding abort circuit 16. concrete
Specifically, the amount of each color component predicted by the quantization width prediction circuit 12
Sign for each color component when actually encoded using child width
The ratio of the code amount to each color component of each standard allocated code amount for each block.
Encoded using the predicted quantization width to match the minute-by-minute ratio.
The amount of code generated for each color component when performing is expressed by the above formula (2),
Predicted from equation (3), and the occurrence mark for each predicted color component.
After allocating the target total amount of code according to the ratio of the amount of code, it is divided into each color component.
Then, based on the statistics for each block, further proportional allocation is performed.
divide In this way, the standard assigned code for each block is
Determine the amount. Note that the ratio of the amount of code for each color component and product block determines the total objective
For example, the method of proportionally distributing the amount of code is
Multiply the code amount by the target total code amount, and multiply it by the total amount of code.
The standard allocation for that block is determined by dividing the code amount of the block.
Decide the amount of code, or divide the target total amount of code into each color component.
5) Proportional distribution with different code amount 5) L, knead each block of its packaging components.
The amount of generated code is distributed proportionally. As a result, the standard allocated code amount for each block is the amount for each color component.
Sign for each color component when actually encoded using child width
The ratio of the code amount to each color component of each standard allocated code amount for each block.
The target total amount of code is distributed so that the ratio of each minute is matched.
This is further distributed proportionally using the predicted code amount for each color component block.
The amount of code is calculated based on the actual amount of code in that block.
A large amount of code is allocated according to the same amount of code. The code amount allocation circuit 20 uses the code amount information table and block allocation.
This code amount data table has a code amount information table.
The code amount information of the corresponding block position is calculated at the code amount calculation time.
While rewriting the input code amount information from the code path 14,
The amount of code for each block input from the code m calculation circuit 14
and the code amount of the entire image, the target total code amount and quantization
Color component identifier given from width prediction circuit 12? 91! Completion
As described above, the coefficient α is calculated for each block for each color component.
Calculate the standard allocated code amount and use the calculated base of each block.
The semi-allocated code amount data is stored in the block allocated code amount data table.
Store it in a bull. This block allocation code amount data table
The standard allocated code amount for each block of the file is
When the variable-length encoding process is performed, the encoding abort circuit 16 -
I can come back. The encoding abort circuit 16 receives the above information from the code amount allocation circuit 20.
One of this block is added to the standard allocated code amount of the corresponding block.
Generated code for allocated code amount in previously processed block
The excess or deficiency amount, which is the difference in amount, is carried over from the previous block.
In addition, the allocated code amount for this corresponding block is
In addition, the code length of each block from the code amount allocation circuit 20
is subtracted from the allocated code amount, and the remaining allocated code amount is sent out.
smaller than the total code amount of power code amount and EOB code.
If the error occurs, a termination signal is output and the variable length encoding circuit 8
, complete the encoding of the block, and
Subtract the transmitted code amount from the allocated code amount and use that value as the allocated code.
It has the function of holding the remainder of the quantity. follow
Then, the encoding abort circuit 16 refers to this allocated code amount,
Sends input code amount to be sent and EOB code
However, if the allocated code amount is not exceeded, no abort is performed.
, complete the naturalization of the block and assign the block
Reduce the amount of code to be sent from the amount of code [and then allocate that chain]
An operation such as holding the remainder of the code amount is performed. Also
, the encoding abort circuit 16 refers to this allocation amount and inputs the
Sends the sign of the input DC component to be sent and the sign of EOB.
If the allocated code amount is not enough to output
After sending out the code of 7, the encoding of that block is terminated.
7, reduce the transmitted code amount from the allocated code amount of the block.
do. In other words, the remainder of the allocated code amount is retained as a negative value.
Then perform the following actions. 10 is a code 81 output circuit, and this n-th output circuit 10 is
Connects the variable length code manually input from the variable length encoding circuit 8.
This is a code that is used to connect the code to a memory card.
The data is written to the recording system 22, which is composed of a recording medium such as a hard drive.
It works like a sea urchin. In this system, each color component determined according to the shooting mode is
Use the commonly used initial standard quantization coefficient α.
Performs J1 processing (first bus) and blocks necessary for optimization.
The amount of information for each color component, the amount of information for each color component, and the information for the entire image.
Investigate the amount, etc., and then, based on the information obtained through this statistical processing,
Starts processing to perform optimized encoding (second bus
). So first block the image, this block
standard amount common to each color component for elements of the image
Quantization using the childization coefficient α, obtained by quantization
Variable length encoding of transform coefficients, and this variable length encoding
The code amount information of each element of each block obtained from
Each color component is calculated from the code amount for each component and the code amount information for the entire image.
Prediction of the coding coefficient α required to obtain the optimal amount of code for each minute.
measurement, proportional allocation of code amount allocated to each color component, and this allocation.
The divided code amount and the code generated in each block of its color components
Determining the standard allocated code amount for each block based on the amount
, a processing model for optimal encoding of the image to be processed based on these.
transition to the
Locking process, for the elements of this blocked image.
quantization using the predicted quantization coefficient α for each color component.
quantization process, the transform coefficients obtained by this quantization [long sign
coding, standard allocated code amount for each block and that block.
The amount of code carried over from the previous processing block is added to the
Variable length encoding processing control within the specified code amount, processing target image
In order to preserve the encoded information of the image, the variable length encoding processing system is
This is called output processing of the variable length code code information obtained by controlling the variable length code.
However, the overall control management is not as shown in the diagram.
It is assumed that the control circuit 18
. Note that this function of the control circuit 18 is performed by a microprocessor.
This can be easily realized by using a processor (MPLI). The above is the configuration of the encoding circuit 80. The recording system 70 in FIG. 1 is an interface circuit 70a.
and a recording medium 7 which is removably connected to this.
Image data encoded and output by the encoding circuit 80
The data and quantization width (or information corresponding to this ≦ this) are
recorded on the recording medium 71 via the interface circuit 70a.
The configuration is as follows. Next, the operation of this device having the configuration shown in FIGS. 1 and 2 above.
I will explain the operation process first, but to get an overview of the whole process.
The basic operation will be explained with reference to FIG. 8 which is a diagram. When using the camera, the user of the camera must switch
30 to set the desired number of shots.
Ru. As a result, the control circuit 90
Find the optimal code amount and use this as the target code amount setting information.
and is applied to the encoding circuit 80. Pictures that can be taken with this model
number is set. When the next photograph is taken, the camera is placed behind the photographic lens 40a.
A subject image is formed as an optical image on the image sensor 40b.
. This image sensor 40b then captures this formed optical image.
is converted into an image signal and output. By the image sensor 'f'-40b, the image signal is
is input to the signal processing circuit 60, where the signal processing circuit 60
Amplification by the amplifier circuit Boa inside, A/D converter 6ob
After exchanging Yoru^/D, temporarily held in buffer Ameseri 6oe
be done. And after this, from Kurahua Memory 60e
The process circuit 60d in the signal processing circuit 60
Accordingly, processing such as band correction and color signal formation is performed. Here, for the later first pass processing, V (luminance phrase), Cr
, in the order of Cb (chroma system, both color difference components) signals.
Therefore, the color signal formation is also affected accordingly. In other words, the image signal is blocked in an 8x8 pixel matrix.
In process circuits, this block is
image (Y component, Cr component from No. 5 data)
(1?4 components), Cb acid component, B-Y component) in the order
In addition to separating the signals of each color component, gamma correction n
Performs white balance processing, etc. 8x8 matrix separated by process circuit God
Image signal of each color component in the blocked image signal of
Data is input to encoding circuit 80. This results in
Image data for one frame (or field)
is divided into blocks of the above-determined size and sequentially
The signal is input to the encoding circuit 80. Furthermore, in the process circuit 60d
The image signals of each color component processed by Y, Cr,
Each color component of Cb is stored in a buffer memory for later processing.
In theory, it may be read out and used. In this embodiment, the image quality (from the No. processing circuit 60) is
)′ component (luminance component) in image signal data.
force is applied, and this is processed later (statistical processing)
After that, all the image data of the Cr component is
Blocking is performed, and statistical processing is performed on this at a later stage.
After that, block the eb acid component image and
The process of performing later statistical processing on
conduct. The encoding circuit 80 receives this signal from the signal processing circuit 60.
Input data is given to the orthogonal transform circuit 4. Then, the orthogonal transformation
The conversion circuit 4 converts input image data into blocks (hereinafter referred to as blocks).
7 for each block (called lock image data)
For example, two-dimensional orthogonal transformation by discrete cosine transform (DCT)
Perform the conversion. This orthogonal transformation by DCT is
Divide the waveform into frequency components and divide this into frequency components as the number of input samples.
Represented by the same number of cosine waves, each with a different frequency
This is the process. Then, orthogonally transformed block image data (transform coefficient
) is an 8×8 matrix in the buffer memory (not shown).
is stored at the corresponding frequency component position on the matrix (matrix).
The origin position of Rix is the DC component, and the rest are AC components.
A relationship in which the frequency increases as the distance from the origin position increases
(stored in a matrix with
6. Then, the quantization circuit 6 converts this block image data (transformed
The first bus (first time) quantization is performed on the coefficient). In this first quantization, each preset frequency component is
every minute (frequency components are for each matrix position of the block)
quantization matrix (determined accordingly) at the time of shooting.
Control circuit 1B corresponds to the image quality setting value set by the user.
The amount multiplied by the standard (provisional) quantization coefficient α given by
Quantize the transform coefficients using the child width (Figure 8 (hl,
i)). The quantization matrix at this time is the luminance system and the chrominance system.
It may be the same for both Ma-kei and Ma-kei, but each
The ability to set a suitable quantization matrix gives good results.
It will be done. The quantized block image data (transform coefficients) is iil
It is input to the variable length naturalization circuit 8, where ri1 variable length n
be converted into In the variable length encoding circuit 8, it is converted into mf and inputted.
Zigzag scan the obtained conversion coefficients in the order shown in Figure 7.
, scans from low frequency components to high frequency components. That is, the transform coefficients are divided into an 8×8 matrix and the frequency
It is stored corresponding to the component, and the closer it is to the meditated point, the higher the frequency.
Since the number is low, zigzag scanning can be used to detect low frequencies.
It is possible to scan from component to higher frequency component. The first data in the scanning order in Figure 7 is the DC component DC.
Therefore, this DC component DC data is immediately preceded by a variable length code.
DC component D of the converted block (previous block)
The difference value djrf-DC with C is manually encoded (
Figure 8 (di), (el,)). For AC components A and C, are they the second in the scanning order in Figure 7?
Look at the conversion coefficients in order from 64th to
When a non-zero (i.e. valid) coefficient appears, immediately
Number of consecutive 0 (invalid) coefficients that existed before (zero run)
) and its effective coefficient values to perform two-dimensional Huffman encoding.
((d2), (e2)). Further, the variable length encoding circuit 8 encodes the 64th coefficient after a certain coefficient.
End of block if invalid coefficients continue up to the coefficient
The EOB (end of block) sign indicating
'Save. (5) The code emitted for that R”y
The length is input H4 to the code amount calculation circuit 14 (gl). This kind of processing for all blocks of one image can be done manually.
I'll go. If this processing for the Y component is completed, then
Next, the same process is applied to each of the Cr and Cb components. On the other hand, the code amount calculation circuit 14 inputs YSCr, C
b Calculate the code amount for the entire image for each color component.
For each block of each component of )゛, Cr, , Cb,
Calculate the code amount and calculate the code amount for each color component and for the entire image.
In addition to performing integration for the body (g2), for each block
The code amount data is output to the code amount allocation circuit 20. sign
The quantity allocation circuit 20 receives this data of η versus quantity for each block.
The sign of the corresponding block position with 1 added to the code amount information table
volume information and

[, write. Then, ゛°US)'' for all blocks of one image,
crscb each generation], at the stage when all human encoding processing is completed, the code amount calculation circuit 14 calculates the data of the code amount for each color component and the code amount for the entire image under the control of the control circuit 18. While outputting the quantization width to the measurement circuit 12,
The code of the entire image] is output to the code amount allocation circuit 20. The quantization width prediction circuit 12 uses the input code amount for each color component, the code amount data for the entire image, and the target total n code amount data to approximate the value of the eye travel code amount by linear prediction, for example. The optimal quantization coefficient α for each color component is predicted based on the jirization coefficients actually used (see Fig. 8).
h2)). Further, the code amount allocation circuit 20 uses the input code amount for each block and the code amount of the entire image, and the predicted quantization coefficient α for each color component given from the eye travel code amount and quantization width prediction circuit 12. The standard allocated code amount of the block is calculated (FIG. 8 (h3)). Specifically, each reference allocated code amount for each block is determined by the ratio of the code amount for each color component when actually encoding using the quantization width for each color component predicted by the quantization width prediction circuit 12. The generated code amount for each color component when encoding is performed using the predicted jilf width is predicted from the above equations (2) and (3) so that the ratio of each color component matches 1. After allocating the target travel amount by the ratio of the generated code amount for each color component,
For each color component, based on the statistics for each block,
Furthermore, it will be distributed proportionately. In this way, the standard allocated code amount for each block is determined. Then, the calculated standard allocated code amount data for each block is stored in the block allocated code amount data table. The data of the allocated code amount for each block in this block allocation + 3 amount data table will be given to the code-naturalization abort circuit j6 when the corresponding block is subjected to variable length encoding processing. This completes the first encoding (statistical processing) for determining the allocated code amount for each block and optimizing the quantization width for the first bus. Next, processing for the second bus begins. This second pass process is
(n-encoding process), and is a process in which the final encoded output optimized to fit within the target total amount is i. This treatment is performed first on the Y component, and after the Y component is completed, the Cr and Cb acid components are treated. The Cr and Cb acid components may be treated in any order. Therefore, if the treatment order is Cr and CbO, the Cr component is treated first, and then the Cb acid component is treated. If the order is Cb and Cr, then the Cb acid component is treated first, and after the Cb acid component has been treated, the Cr component is treated. In the 2 best 1 process, image data is first read out in blocks, and image transmission data of the r component (luminance system) (which is extracted and output from the signal processing circuit 60) is manually input to the encoding circuit 80. (a). The input blocked image data is input to the orthogonal transformation circuit 4 in the encoding circuit 80, and orthogonal transformation is performed again (b). The transform coefficients obtained by this orthogonal transform are input to the quantization circuit 6,
Quantization is performed again (C). However, [5, the quantization coefficient a used at this time is the optimum quantization coefficient a of the prediction for each color component calculated by the quantization width prediction circuit 12 in the previous bus.
It is. Next, the transform coefficients of the quantized block image data are input to the variable length encoding circuit 8. In variable length encoding, as in the conventional processing, among the transform coefficients of this block image data, first the difference value diff-DC of the DC component I)C is Huffman encoded (di), (el)), and then Data is extracted sequentially from the alternating current component AC by zigzag scanning.
Perform dimensional Huffman encoding ((d2). (e2)). However, each time a Huffman code for one element (one position in the matrix) is generated, the code amount allocation circuit 20 sends the standard allocated code amount for that block stored in the allocated code amount data table to the encoding abort circuit. On the other hand, the encoding abort circuit 16 adds the remainder to the reference assigned code of each block and the assigned code in the encoding process of the previous block to obtain the assigned code amount of the corresponding block, and sends it out. If the allocated code amount is not exceeded even if the exponent code amount and the IEOB code are sent, a process is performed to reduce the code amount to be sent from the allocated code amount of the block without generating an abort signal. Then, when the total code amount of the code amount of the block to be sent and the EOB code exceeds the remaining code amount of the allocated code amount, the encoding abort circuit i6 starts the variable length encoding circuit. 8, a termination signal is output to terminate the Huffman encoding of the block, and the transmitted code amount is subtracted from the allocated code amount of the block, and the value is held as the remainder of the allocated code amount. The variable length encoding circuit 8 then proceeds to Huffman encoding the next block obtained from the quantization circuit 6. Further, in the encoding aborting circuit 16, the DC component and the EOB
If the amount of allocated codes is insufficient to send out the codes, these codes are sent out and then sent to the coding abort circuit 16.
outputs an abort signal to the variable length encoding circuit 8 to end the Huffman encoding of the block, and also subtracts the transmitted code amount from the allocated code amount of the block and sets the remainder of the allocated code amount as a negative value. Hold. Therefore, the "J variable length encoding circuit 8 is the encoding abort circuit 1 [i
The converted '\huma', - code is output to the code output circuit 10 until a truncation signal is input from , and if Huffman encoding is completed for all elements of the matrix before the truncation signal is generated, The variable length encoding circuit 8 is EO
The code of B is output to the code output circuit 1o. Furthermore, if the variable length encoding circuit 8 receives an abort signal before completing Huffman encoding for all elements of the matrix, the code output circuit 1o outputs the EOB code instead of that code.
It will be output to . The encoded data is temporarily stored in the code output circuit 1o. The variable length encoding circuit 8 then proceeds to Huffman encoding of the next block obtained from the quantization circuit 6. By repeating such operations and completing processing of all blocks of one screen image, all encoding processing is completed. Once the Y component is processed in this way, the chroma components (Cr, Cb) are processed in a similar manner. In processing this second component as well, the quantization circuit 6 uses the T'-measured optimal sonization coefficient α for the chroma component calculated by the quantization width 'f-measurement circuit 12 in the previous bus. As shown in step 2 above, all the encoding processes are completed by completing the processing of the second bus for all blocks of one screen's worth of images for all color components. Upon completion of this process, the encoder output circuit 10 outputs the optimized Huffman encoded data for the image to the recording system 22, where it is stored in a storage medium 7 such as a memory card.
] (f). This is performed by the output of the code output circuit 10, which connects the Huffman code, which is a variable length code, from the variable-length encoding circuit 8 and provides it to the memory card, which is the storage medium 71. Write about it. The writing to the storage medium 71 by the output of the code output circuit 10 may be performed all at once when the second bus is completed, or it may be performed once the first bus is completed and the second bus is executed. The results of connecting variable-length Huffman codes may be sequentially written to the storage medium in units of one byte or several bytes as soon as they are assembled. Note that, prior to this, the code output circuit 10 writes information on the quantization matrix of each color component used for encoding into the header part of the stored data of the encoded image, and uses it as a clue during playback and 1. I'll leave it there. As described above, in this apparatus, statistical processing is first performed, and through this statistical processing, the necessary techniques are used to determine the predicted value of the optimal quantization width for each color component and the base map allocation code amount for each color component-specific product block. Based on this information, predict the optimal quantization width for each color component, allocate the code amount for each color component,
After determining the standard allocated code amount for each color component block using this allocated code amount and information on the generated code amount for each color component block, the actual encoding process begins. Separately, look at the encoded output for each block in that color component in turn and assign it to that block.
The allocated code amount for that block, which is obtained by combining the standard allocated code amount calculated by C and the excess (carryover) of the actual output code amount for the allocated code amount in the processing block immediately before the block. The amount of code for each block is controlled by terminating the variable length encoding so that the amount of code is within the desired amount, and the final output is a coded output that is within the desired amount of code. This point is an important point of the present invention. Therefore, the block size, the type of orthogonal transformation, the type of variable length encoding, etc. used in this embodiment are not limited. Also, statistical processing does not necessarily have to be done once;
You don't have to do it, or you can do it a few more times. However, it is best to perform statistical processing once in terms of processing time and effects obtained. In addition, statistical processing does not necessarily require actual encoding to check the generated code amount; instead, it is a method that calculates the activity for each block and calculates the standard allocated code amount and quantization coefficient for each block from the results. It's okay to have one. In addition, it is not necessarily necessary to allocate all of the code amount of the memory to each block. For example, when the code amount of the memory is allocated to the allocated code amount of each block, the remainder may be used for the first block and the previous processing block. In addition, in this embodiment, the EOB code amount is included in the determination of the standard allocated code amount for each block and the code amount control. Since it does not change even if the quantization width changes, it is of course possible to exclude this EOB code amount in advance and determine the standard allocated code amount for each block and control the code amount. It is desirable because it improves Similarly, in predicting the quantization width, the EOB code amount is excluded. It is better to perform prediction using the code amount. Also, the compression ratio does not have to be variable, it is fixed to one type of compression ratio, and the objective total summation coefficient, provisional quantum coefficient, etc. may all be set as fixed values. The configuration becomes simpler. Further, the image data buffer memory may be provided between the orthogonal transform circuit 4 and the quantization circuit 6, and in this case, the blocking and orthogonal transform processes in the encoding process can be omitted. However, in this case, the size of the image memory increases to maintain accuracy. Also, process processing is performed before A/D conversion.
You may digitize it afterwards. In addition, in this device, variable length encoding is performed for each block starting with low frequency components, and high frequency components that have relatively little visual impact on image quality are omitted by discontinuing encoding. Encoding can be performed with high compression while minimizing deterioration in image quality. The present invention having the configuration shown in FIGS. 1 and 2 described above has the following features:
In short, based on the results of statistical processing, we predict the quantization coefficient for each color component, and also compare the amount of code for each color component when actually encoding using the quantization width for each color component. To,
The target travel pair amount is distributed so that the ratio of each standard allocated code amount for each color component is matched, and this is further distributed proportionally by the predicted energy amount for each color component product block. The standard allocated code amount for the block is determined, and in the encoding process, quantization and variable length encoding of each block for each color component using the F11 Nf coefficient for each color component, and variable length encoding are performed. The amount of code obtained by the variable-length encoding is made to fit within the allocated code amount obtained by adding the standard allocated code amount of the block to be encoded and the excess or deficiency of the allocated code amount in the processing block before that block. , proceed with variable length encoding while monitoring the code amount for each block, and when the code amount reaches the allocated code amount, the coding of that block is terminated, or even if the DC component exceeds the allocated code amount, it is necessary to Encode it and store the excess or deficiency of the allocated code amount at that time5.
Then, the next block is encoded. This means that if you simply use the allocated code amount for each block, which is obtained by proportionally allocating the target code amount by the amount of code for each block, for a certain color component, the allocated code amount for each block or the extra Then, if there is a shortage in a certain color component, the target amount of travel code is allocated based on the predicted ratio of the emitted code amount for each color component, and this is further calculated based on the statistics for each block. By calculating the allocated code amount for each block, we can calculate the ratio of sign + 3 amount for each color component and the generated code amount when actually n-coding. Since the ratio is very close to the ratio for each inclusive phrase, it is possible to make the allocated code amount for each block reasonable.By making such a rational allocation in advance, each color component This is to enable balanced encoding. Therefore, according to the present invention, it becomes possible to encode images with a wide range of spatial frequencies within a certain amount of code within a certain processing time without degrading the image quality. In compression encoding, the encoding circuit 80 having the configuration shown in FIG. , based on the results, find the optimal quantization width and calculate this optimal ffi
The process is completed in two steps: the second pass is carried out depending on the f-size width, and the final compressed and encoded data is obtained.The first pass is used to determine the optimal a for each color component and the This is to determine the standard allocated activity for each product block. Figure 1 3: Process flow of the encoding circuit 80. In FIG. 3, in order to clearly understand the flow of processing in the encoding circuit 80, the flow of signals on the first bus is indicated by dotted arrows, and the flow of signals on the second bus is indicated by solid arrows. 7, respectively. If you follow the operation along the flow of this signal, it will look like the following. When encoding image data, a target travel pair quantity is set in the control circuit 18 of the encoding circuit 80. By operating the switch 30, the user sets the desired number of shots that can be taken.
This is achieved by the control circuit 90 selecting the optimum code amount according to the number of sheets of information, and providing this to the encoding circuit 80 as information on the number of travel code amounts. Note that in the initial state, the number of images that can be taken is set to a predetermined standard number. When photographing is performed, this causes the imaging system f in the imaging system 40 to
An image signal is output from. This outputted image signal is converted into a digital signal in the signal processing circuit 60, stored in a buffer memory, and then read out in blocks of 8x8 pixels.
The acid component is separated. This separation is first performed on the )′ component, and the Y component image data is output in blocks of 8×8 pixels.
In this example, I) CT; Discrete cosine transform (Disc,
rcl, e Co51ne Transrorm) is performed to decompose the data into frequency components. The DCT transform coefficients obtained by the orthogonal transform circuit 4 are input to the quantization circuit 6 , while the control circuit 18 outputs the number of travel codes to the quantization width prediction circuit 12 .
In step 2, the initial value of the quantization coefficient α is set from the travel code amount using the relationship of equation (1), and is output to the quantization circuit 6. In the quantization circuit 6, the quantization coefficient α(
That is, the transform coefficients are linearly quantized using the provisional quantization coefficient α). The quantized transform coefficients are input to a variable length encoding circuit 8, and variable length encoding (Huffman encoding in this example) is performed. The quantization coefficients input here are scanned from low frequency components to high frequency components called jigsag scanning, and the data of the first DC component is the DC component of the block that was variable-length encoded immediately before. The difference value is Huffman encoded and output. For the AC component, look at the conversion coefficients in order from the 2nd to the 64th in the scanning order, and when a conversion coefficient or a non-zero (in other words, valid) coefficient appears, it is , invalid) coefficients (zero run) and their effective coefficients, two-dimensional Huffman encoding is performed. Also, if invalid output continues after a certain coefficient up to the 64th coefficient, EOB indicating the end of the block will be displayed.
Outputs the sign of . The variable length improvement circuit 8 performs encoding as described above,
Every time one code occurs, its code length is output to the code amount calculation circuit 14. The code amount calculation circuit 14 stores each input code length for each color component block, and also accumulates and stores the code lengths for each color component. 〜゛When this kind of processing for the component is finished, next C
r component,

[7] Similar processing is performed for the cb component. When the encoding of one image is completed, the code amount calculation circuit 14 adds the accumulated code amounts for each block for each color component and calculates the code amount for each color component and the code amount for the entire image as a travel value. do. The code amount and travel pair value for each color component are output to the quantization width prediction circuit M+, 2, and are also output to the code amount for each color component and the code amount allocation circuit 20 for the entire image. When the encoding process of the first bus is completed, the code amount allocation circuit 20 receives the code amount for each color component inputted from the code amount calculation circuit 14, the code amount for the entire image, the target code amount, and the code amount for each color component. Based on the code amount of the block, a standard allocated code amount for each block for each color component is calculated and stored by proportional allocation. In addition, the quantization width prediction circuit 12 calculates a more suitable quantization coefficient α for each color from the picture image code amount obtained by encoding on the first bus and the “J travel code amount” given from the control circuit 18. The predictions 1-7 are output for each component to the quantization circuit 6.The code amount allocation circuit 20 also outputs the prediction quantization coefficient α for each color component given from the quantization width f-measurement circuit 12. The standard allocated code amount for each block is calculated from the above. Subsequently, encoding processing on the second bus is performed on the same image data. On the second bus, the image data read from the memory in the signal processing circuit 60 is separated into color components in the order of processing on the first bus, and the image data of each component is divided into 8×
After processing such as 8-pixel blocking is performed, the data is input to the orthogonal transform circuit 4, and orthogonal transform (DCT transform) is performed for each block in order.
The T transform coefficients are input to the quantization circuit 6. On the other hand, the quantization circuit 6 linearly quantizes the DCT transform coefficients using the corrected quantization width based on the new quantization coefficient α for each color component based on the given prediction. The quantized coefficients are input to the variable length encoding circuit 8, and the first
Huffman encoding is performed using the same method used for bus encoding. Here, the amount of code generated during encoding is obtained at the end of the first bus, and processing is performed on the corresponding sign of the standard allocated code amount of each color component block stored in the code amount allocation circuit 20. The assigned code of the block is the sum of the standard assigned code I of the block and the excess/deficiency of the code amount (carryover) with respect to the assigned code amount in the encoding processing block immediately before the block to be encoded. Comparison is performed using the amount as a reference value, and when this reference value is exceeded, the subsequent coding within that block is aborted by the action of the code abort circuit 16. Furthermore, the excess or deficiency of the allocated code amount at the time when the encoding of the block is finished is stored in the code amount allocation circuit 20. The encoded data controlled to the target edge code amount by the above method is sequentially outputted to the recording system 70 via the code output circuit 10 and recorded. Next, reproduction of compressed and encoded recorded image data recorded on the recording medium 71 by the recording system 70 will be explained. Figure 4 shows the configuration of the playback machine. In the figure, 100 is the main body of the playback machine, and this main body 100 of the playback machine reads 71B3.
02, decoding circuit] 04, a processing circuit 106, and a control circuit 108. Reading unit]02 can attach and detach the storage medium 71, and reads the contents of the storage medium 71 to the interface circuit 1.
10. The decoding circuit 104 has functional blocks as shown in FIG. That is, ]12 is a Huffman decoding unit that decodes Huffman encoded data, and 11'4 is a Huffman decoding unit that reads out the data obtained by the /% Huffman decoding from the storage medium 71 and converts it to the input quantization width. An inverse quantization unit 116 performs inverse quantization based on information, and an I inverse DCT transforms the data obtained by inverse quantization and outputs it as video signal data.
A DcT (inverse DCT transform) section, and 118 a control section that controls these. Processing time WHOs includes buffer memory 120 and encoder 1
22 and a D/H converter 124. The buffer memory 120 is a memory that temporarily holds the video signal data output from the decoding circuit 104, and the encoder 122 converts the video signal data read out from the buffer memory 120 into an NTSC' video signal. can be,
The D/C converter 124 converts this NTSC video signal into analog and outputs it as a video signal for television. The control circuit 108 is in charge of overall control of the player main body 100, and controls the reading section 102 of the player main body 100 to read out information on the quantization width at the time of encoding.
As a result, the information on the quantization width at the time of encoding read from the recording medium 71 is set in the inverse quantization unit 114 of the decoding circuit 104, and then the control circuit 108 sets the information of the quantization width at the time of encoding read from the information recording medium 71. In order to read out the video signal data, the reading section 102 is controlled. Further, although not shown, the playback main body 100 includes a frame advance switch and the like, and the video at the frame position specified by this switch can be played back. The control circuit 108 also performs such control. Next, the operation of the reproducing machine having the above configuration will be explained. When the recording medium (memory card) 71 on which compression-encoded video signal data is recorded is attached to the reading section 102 of the main body 100 of the playback device, the control circuit 108 first tells the reading section 102 to
Control is performed to read out information on the quantization width for each color component during encoding. As a result, the reading unit 102 reads the recording medium 7.
Information about the quantization width at the time of encoding is read from I, and this information is set in the inverse quantization section 1.14 of the decoding circuit 104. Next, the control circuit 108 controls the reading section 102 to read the video signal from the recording medium 71, so that the reading section 1
02 sequentially reads video signals from the recording medium 71 and inputs them to the decoding circuit 104. In the decoding circuit 104 that receives this, the Huffman decoding unit 1
12, the Huffman code is decoded to obtain quantized coefficients. The quantized coefficients obtained in this way are supplied to an inverse quantization circuit 114 for inverse quantization. The inverse quantization here is performed using the previously set information on the quantization width for each color component. The transform coefficients obtained by inverse quantization are transferred to the IDCT section 11
In B, each block is subjected to inverse DCT transformation and restored to the original video signal. In this way Y, Cr. The video signals are restored in CbO order, output from the decoding circuit 104, and written into the buffer memory 120 in the processing circuit 10B. When the writing of one screen of video signal data is completed, the video signal data is read out from the buffer memory 120 in the normal scanning order of the television signal, and the encoder 1
At step 22, the NTSC video (≦) signal is converted. Further, it is converted into an analog signal by an l/Δ converter 124, and output. By inputting this video (, -) to a TV monitor, it will be played back as an image or TV video,
It can be viewed as a video, and a hard copy can be obtained by printing with a printing device such as a video printer, so it can be viewed in the same way as a photograph. As explained above, the camera can set the desired number of shots that can be taken, and by setting the number of shots that can be taken, the camera automatically sets the corresponding compression rate, and the compression rate is initially determined according to this set compression rate. Using the provisional quantization width,
]Screen's worth of photographed image data is sequentially quantized and variable-length coded for each color component in each block, and from the code amount of each color component and screen's worth of photographed image data obtained as a result, the "optimum for each color component" is determined. fLF conversion width" is measured, and the code amount that can be assigned to each color component is determined by proportional allocation from the code amount for each color component, and the standard code amount to be allocated to each block for each color component is determined by that block. Then, using the predicted optimal quantization width for each color component, the captured image data for one screen is sequentially quantized again for each block for each color component. , when performing variable-length encoding on this quantized output, the standard allocated code amount for each block and the remainder of the allocated code amount during encoding in the encoding process of the block processed one before it are added. Variable length encoding is performed within this range with the allocated code amount as the maximum frame, and when the encoded video signal is played back, it is decoded using the optimum quantization width used at the time of shooting, thereby achieving the desired compression rate. can be encoded without installing hardware for different compression rates.
It can be realized with one common hardware, and similarly, decoding of image data encoded with a desired compression rate can be realized with one common hardware without providing hardware for each compression rate. As described above, in the embodiment, the encoding process or the first bus (statistical processing)
] times, and the second bus (encoding process) times, which completes the process two times in total. However, the method is not limited to this, and it will be even better if the first bus is repeated several times. results. Further, although an example is shown in which a memory card is used as the recording medium, other media such as a floppy disk, an optical disk, and a magnetic tape may also be used. Further, although the camera and the playback device are shown as separate bodies, it is also possible to use an integrated camera in which the camera also has the functions of the playback device. Furthermore,
In the embodiment, the quantization width value itself is recorded on the recording medium, but the quantization width value may be converted or encoded and then recorded. In addition, in the above example, the quantization width was set based on the amount of code for the travel code, but in applications where multiple code amounts for the travel code are switched and used by mode, the quantization width corresponding to each mode is Of course, it is also possible to prepare widths in advance and use them by switching them depending on the mode. It should be noted that the present invention is not limited to the embodiments described above and shown in the drawings, and can be implemented with appropriate modifications within the scope of the gist thereof.For example, the present invention is not limited to still images. It can be applied to compression encoding of various images such as moving images. In addition, in the above embodiment, information on travel vs. amount is given from compression ratio correspondence information, and a quantization coefficient for each color component that can give a quantization width corresponding to this travel vs. amount is predicted. Quantization is performed using the quantization width based on the quantization coefficient, but the relationship between the quantization width and the compression rate correspondence information is calculated in advance and created in a table. This is stored in memory etc. and the compression It is also possible to output quantization width information (ie, quantization coefficient or quantization width value) directly from the rate correspondence information. In this way, when information corresponding to the desired compression ratio is input, information on the quantization width that is uniquely determined according to the input compression ratio information can be read and output, and optimal quantization can be performed immediately. Immediately after setting the width, the quantization circuit can quantize the image signal data using this quantization width. According to this configuration, information on the optimal quantization width corresponding to various compression ratios is stored in advance in a memory, etc., and the information on the optimal quantization width is stored in correspondence with the input compression ratio information without being repeatedly displayed. Therefore, when encoding to fit within the code amount of the code, the processing can be done in an extremely short time and the hardware can be simple. [Effects of the Invention] As detailed above, according to the present invention, it is possible to control the amount of code for images with a wide range of spatial frequencies without degrading the image quality, and it is possible to control the amount of code within a certain amount of code in a certain processing time. It is possible to provide an image data encoding device and encoding method that are optimal for electronic camera systems, etc., and can perform encoding processing so that the image data fits within the range of .

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing a first embodiment of the present invention;
The figure is a block diagram showing an example of the main part configuration of the device of the present invention, FIG. 3 is a block diagram for explaining the flow of operation of the circuit shown in FIG. 2, and FIGS. 4 and 5 are block diagrams showing the configuration of the regenerator. 6 is a perspective view showing the external appearance of the electronic camera body according to the present invention, FIG. 7 is a diagram illustrating zigzag scanning of a block divided into 8×8 pixels, and FIG. 8 is a diagram showing the principle of the present invention. FIG. C9 is an operational transition diagram for explaining the conventional technique, and FIG. 10 is a diagram for explaining the relationship between code amount and quantization width. 1... Electronic camera body, 6... Quantization circuit, 8...
Variable length encoding circuit, ]O... - code output circuit, 12. quantization width and #1 circuit, 14... code amount calculation circuit, 1B.
...? No. 1 partition termination circuit, 18.18a...control circuit, 2
0... Code amount allocation circuit, 24... DCPM circuit, 3
0...Switch, 40...Imaging thread, 48...1. ,,C
I) Display device, 60...signal processing circuit, 80...encoding circuit, 70...recording system, 71...recording medium, 90 control circuit, JOD...player main body, 102...reading section , 104...Decoding circuit, lO6... Processing circuit, J
H... Control circuit, 110 - Interface circuit, 11
2... Huffman decoding unit, 114 - Inverse quantity conversion unit, 11
6...IDcT (inverse +1) CT conversion) section, 118...
・Control unit. Applicant's agent Patent attorney - 7 Atsushi Tsuboi Figure 6 Figure 15 Figure 7 ogSF Quantized f series --- 1 -- Figure 10

Claims (17)

[Claims] (1) Divide each of the color image data separated by color component into blocks of a predetermined number of pixels, perform orthogonal transformation for each divided block by color component, and then quantize the transformed output by quantization means. Then, this quantized output is given to a variable-length encoding means to be variable-length encoded, and when encoding image data, it is first quantized with a provisionally determined quantization width and then variable-length encoded. , performs statistical processing to obtain the code amount for each block, and then calculates the code amount for each block obtained through this statistical processing and the total code amount target value per image to find the appropriate amount to fit within this total code amount target value. The optimal quantization width and allocated code amount for each block are determined, and the transform output for each block is quantized according to the determined optimal quantization width, and the quantization output is quantized for each block. In an encoding device that sequentially performs variable length coding while controlling the code amount for each block according to the allocated code amount, each color component is Calculating means for calculating the code amount of each block; and predicting the code amount for each color component when quantized with the optimum quantization width, and based on the predicted code amount for each color component, the total code amount target. a determining means for proportionally allocating the value and determining a code amount for each color component, and also for proportionally allocating this according to the calculated value of the calculation means to determine a reference code amount for each block for each color component; Using the reference code amount for each block for each color component determined by the determining means, the code amount generated by encoding in the processed block processed before the next target block to be encoded among the blocks is calculated. Allocation means that allocates the code amount obtained by calculating the difference from the allocated code amount in the processed block and adding it to the reference code amount of the processing target block as the allocated code amount in the processing target block; Encoding of image data, characterized by comprising: aborting means for aborting the variable length encoding so as not to exceed the allocated code amount allocated by the allocating means in the encoding process of the processing target block. Device. (2) Divide each of the color image data separated by color component into blocks of a predetermined number of pixels, perform orthogonal transformation for each divided block by color component, and then quantize the transform output by quantization means. Then, this quantized output is given to a variable-length encoding means to be variable-length encoded, and when encoding image data, it is first quantized with a provisionally determined quantization width and then variable-length encoded. , performs statistical processing to obtain the code amount for each block, and then calculates the code amount for each block obtained through this statistical processing and the total code amount target value per image to find the appropriate amount to fit within this total code amount target value. The optimal quantization width and allocated code amount for each block are determined, and the transform output for each block is quantized according to the determined optimal quantization width, and the quantized output is allocated for each block. In an encoding device that performs a coding process of sequentially performing variable length coding while controlling the code amount for each block according to the code amount, each block is coded for each color component based on the variable length coding output during the statistical processing. a calculation means for calculating a code amount for each color component when quantized with the optimum quantization width, and proportionately allocating the total code amount target value based on the predicted code amount for each color component; a determining means for determining a code amount for each color component and proportionally distributing it according to the calculated value of the calculation means to determine a reference code amount for each block for each color component; Using the determined standard code amount for each block for each color component, calculate the code amount generated by encoding in the processed block processed before the next target block to be encoded among the blocks and the processed block. an allocation unit that allocates a code amount obtained by calculating the difference between the code amount and the code amount allocated to the block to be processed and adding it to the reference code amount of the block to be processed as the amount of code to be allocated to the block to be processed; terminating means for aborting the variable length encoding so as not to exceed the allocated code amount allocated by the allocating means in the encoding process; 1. An image data encoding device comprising: encoding output means for outputting as a code. (3) Divide each of the color image data separated by color component into blocks of a predetermined number of pixels, perform orthogonal transformation for each divided block by color component, and then quantize the transformed output by quantization means. Then, this quantized output is given to a variable-length encoding means to be variable-length encoded, and when encoding image data, it is first quantized with a provisionally determined quantization width and then variable-length encoded. , performs statistical processing to obtain the code amount for each block, and then calculates the code amount for each block obtained through this statistical processing and the total code amount target value per image to find the appropriate amount to fit within this total code amount target value. The optimal quantization width and allocated code amount for each block are determined, and the transform output for each block is quantized according to the determined optimal quantization width, and the quantized output is allocated for each block. In an encoding device that performs a coding process of sequentially performing variable length coding while controlling the code amount for each block according to the code amount, each block is coded for each color component based on the variable length coding output during the statistical processing. calculation means for calculating the total code amount for each color component and the total code amount for the entire image; quantization width prediction means for predicting a quantization width; and predicting a code amount for each color component when quantized with the optimum quantization width, and determining the total code amount target value based on the predicted code amount for each color component. is distributed proportionally, the code amount for each color component is determined, and the standard code amount for each block for each color component is determined by distributing this according to the code amount for each block for each color component calculated by the calculation means. and determining means for determining a processed block processed before a target block to be encoded next among the blocks, using a reference code amount for each block for each color component determined by the determining means during encoding processing. The code amount obtained by calculating the difference between the code amount generated by the encoding and the allocated code amount in the processed block and adding this to the reference code amount of the processing target block is calculated in the processing target block. an allocation means for allocating an allocated code amount as an allocated code amount, and an abort means for terminating the variable length encoding so as not to exceed the allocated code amount allocated by the allocation means in the encoding process of the processing target block. An image data encoding device characterized by: (4) Divide each of the color image data separated by color component into blocks of a predetermined number of pixels, sequentially perform orthogonal transformation for each divided block by color component, and then quantize the transformed output by quantization means. Then, this quantized output is given to a variable-length encoding means to be variable-length encoded, and when encoding image data, it is first quantized with a provisionally determined quantization width and then variable-length encoded. , performs statistical processing to obtain the code amount for each block, and then calculates the code amount for each block obtained through this statistical processing and the total code amount target value per image to find the appropriate amount to fit within this total code amount target value. The optimal quantization width and allocated code amount for each block are determined, and the transform output for each block is quantized according to the determined optimal quantization width, and the quantized output is allocated for each block. In an encoding device that performs a coding process of sequentially performing variable length coding while controlling the code amount for each block according to the code amount, each block is coded for each color component based on the variable length coding output during the statistical processing. a calculation means for calculating a code amount for each color component when quantized with the optimum quantization width, and proportionately allocating the total code amount target value based on the predicted code amount for each color component; a determining means for determining a code amount for each color component and proportionally distributing it according to the calculated value of the calculation means to determine a reference code amount for each block for each color component; Using the determined standard code amount for each block for each color component, calculate the code amount generated by encoding in the processed block processed before the next target block to be encoded among the blocks and the processed block. an allocation unit that allocates a code amount obtained by calculating the difference between the code amount and the code amount allocated to the block to be processed and adding it to the reference code amount of the block to be processed as the amount of code to be allocated to the block to be processed; terminating means for aborting the variable length encoding so as not to exceed the allocated code amount allocated by the allocating means in the encoding process; An image data encoding device comprising: encoding output means for outputting as a code; and encoding output means for outputting the transform coefficients variable-length encoded by the variable-length encoding means as a code. (5) Divide each color image data separated by color component into blocks of a predetermined number of pixels, sequentially perform orthogonal transformation for each divided block by color component, and then quantize the transform output by quantization means. Then, this quantized output is given to a variable-length encoding means to be variable-length encoded, and when encoding image data, it is first quantized with a provisionally determined quantization width and then variable-length encoded. , performs statistical processing to obtain the code amount for each block, and then calculates the code amount for each block obtained through this statistical processing and the total code amount target value per image to find the appropriate amount to fit within this total code amount target value. The optimal quantization width and allocated code amount for each block are determined, and the transform output for each block is quantized according to the determined optimal quantization width, and the quantized output is allocated for each block. In an encoding device that performs sequential variable length coding while controlling the code amount for each block according to the code amount, the target compression rate is selected and the data should be stored according to the selected compression rate. Compression rate selection means for determining the target total code amount and the provisional quantization width; Calculation means for calculating the code amount for each block for each color component based on the variable length encoded output during the statistical processing; , predicting the code amount for each color component when quantized with the optimum quantization width, and proportionally distributing the total code amount target value based on the predicted code amount for each color component, and determining the code amount for each color component. and determining means for proportionally distributing this in accordance with the calculated value of the calculating means to determine a standard code amount for each block for each color component, and a standard for each block for each color component determined by the determining means during encoding processing. Using the code amount, calculate the difference between the code amount generated by encoding in the processed block processed before the next block to be encoded among the blocks and the allocated code amount in the processed block. an allocation means for allocating a code amount obtained by calculating the code amount and adding it to the reference code amount of the processing target block as an allocated code amount for the processing target block; and an allocation means in the encoding process for the processing target block. an aborting means for aborting the variable length encoding so as not to exceed the allocated code amount allocated by the image data encoding apparatus. (6) Divide each of the color image data separated by color component into blocks of a predetermined number of pixels, perform orthogonal transformation for each divided block by color component, and then quantize the transformed output by quantization means. Then, this quantized output is given to a variable-length encoding means to be variable-length encoded, and when encoding image data, it is first quantized with a provisionally determined quantization width and then variable-length encoded. , performs statistical processing to obtain the code amount for each block, and then calculates the code amount for each block obtained through this statistical processing and the total code amount target value per image to find the appropriate amount to fit within this total code amount target value. The optimal quantization width and allocated code amount for each block are determined, and the transform output for each block is quantized according to the determined optimal quantization width, and the quantized output is allocated for each block. In an encoding device that performs sequential variable length coding while controlling the code amount for each block according to the code amount, the target compression rate is selected and the data should be stored according to the selected compression rate. Compression rate selection means for determining the target total code amount and the provisional quantization width; Calculation means for calculating the code amount for each block for each color component based on the variable length encoded output during the statistical processing; , predicting the code amount for each color component when quantized with the optimum quantization width, and proportionally distributing the total code amount target value based on the predicted code amount for each color component, and determining the code amount for each color component. and determining means for proportionally distributing this in accordance with the calculated value of the calculating means to determine a standard code amount for each block for each color component, and a standard for each block for each color component determined by the determining means during encoding processing. Using the code amount, calculate the difference between the code amount generated by encoding in the processed block processed before the next block to be encoded among the blocks and the allocated code amount in the processed block. an allocation means for allocating a code amount obtained by calculating the code amount and adding it to the reference code amount of the processing target block as an allocated code amount for the processing target block; and an allocation means in the encoding process for the processing target block. terminating means for terminating the variable length encoding so as not to exceed the allocated code amount allocated by the variable length encoding means; and encoding output means for outputting the transform coefficients variable length encoded by the variable length encoding means as codes. An image data encoding device comprising: (7) Divide each of the color image data separated by color component into blocks of a predetermined number of pixels, sequentially perform orthogonal transformation for each divided block by color component, and then quantize the transformed output by quantization means. Then, this quantized output is given to a variable-length encoding means to be variable-length encoded, and when encoding image data, it is first quantized with a provisionally determined quantization width and then variable-length encoded. , performs statistical processing to obtain the code amount for each block, and then calculates the code amount for each block obtained through this statistical processing and the total code amount target value per image to find the appropriate amount to fit within this total code amount target value. The optimal quantization width and allocated code amount for each block are determined, and the transform output for each block is quantized according to the determined optimal quantization width, and the quantized output is allocated for each block. In an encoding device that performs sequential variable length coding while controlling the code amount for each block according to the code amount, the target compression rate is selected and the data should be stored according to the selected compression rate. compression rate selection means for determining the target total code amount and the provisional quantization width; a calculation means for calculating the total code amount for each component and the total code amount for the entire image; and a quantization width for predicting the optimal quantization width for each color component based on the total code amount target value and the calculated value by the calculation means. a prediction means, predicting a code amount for each color component when quantized with the optimum quantization width, proportionately allocating the total code amount target value based on the predicted code amount for each color component, and calculating a code amount for each color component; determining means for determining a reference code amount for each block for each color component by determining the amount and allocating this according to the code amount for each block for each color component calculated by the calculation means; Using the standard code amount of each block for each color component determined by the determining means, the amount of code generated by encoding in the processed block processed before the target block to be encoded next among the blocks. and allocation means for allocating the code amount obtained by calculating the difference between the code amount and the allocated code amount for the processed block and adding this to the reference code amount for the processing target block as the allocated code amount for the processing target block; , a terminating means for terminating the variable length encoding so as not to exceed the allocated code amount allocated by the allocating means in the encoding process for the processing target block. conversion device. (8) Divide each of the color image data separated by color component into blocks of a predetermined number of pixels, perform orthogonal transformation for each divided block by color component, and then quantize the transform output by quantization means. Then, this quantized output is given to a variable-length encoding means to be variable-length encoded, and when encoding image data, it is first quantized with a provisionally determined quantization width and then variable-length encoded. , performs statistical processing to obtain the code amount for each block, and then calculates the code amount for each block obtained through this statistical processing and the total code amount target value per image to find the appropriate amount to fit within this total code amount target value. The optimal quantization width and allocated code amount for each block are determined, and the transform output for each block is quantized according to the determined optimal quantization width, and the quantized output is allocated for each block. In an encoding device that performs sequential variable length coding while controlling the code amount for each block according to the code amount, the target compression rate is selected and the data should be stored according to the selected compression rate. Compression rate selection means for determining the target total code amount and the provisional quantization width; Calculation means for calculating the code amount for each block for each color component based on the variable length encoded output during the statistical processing; , predicting the code amount for each color component when quantized with the optimum quantization width, and proportionally distributing the total code amount target value based on the predicted code amount for each color component, and determining the code amount for each color component. and determining means for proportionally distributing this in accordance with the calculated value of the calculating means to determine a standard code amount for each block for each color component, and a standard for each block for each color component determined by the determining means during encoding processing. Using the code amount, calculate the difference between the code amount generated by encoding in the processed block processed before the next block to be encoded among the blocks and the allocated code amount in the processed block. an allocation means for allocating a code amount obtained by calculating the code amount and adding it to the reference code amount of the processing target block as an allocated code amount for the processing target block; and an allocation means in the encoding process for the processing target block. terminating means for terminating the variable length encoding so as not to exceed the allocated code amount allocated by the variable length encoding means; and encoding output means for outputting the transform coefficients variable length encoded by the variable length encoding means as codes. An image data encoding device comprising: . (9) The quantization means performs quantization during the encoding process from data of low frequency components among the frequency component data obtained by the orthogonal transformation, and
The variable-length encoding means obtains the difference between the DC component and the DC component in the previous processing block among the frequency component data of the image data for each block obtained by the orthogonal transformation for each frequency component. , a DC component encoding processing unit that variable-length encodes this difference, and an AC component encoding processing unit that performs variable-length encoding on the AC component after the encoding processing unit finishes processing the DC component. 9. The image data encoding apparatus according to claim 1, wherein the encoding aborting means is provided only in the AC component encoding processing section. (10) Divide the color image data separated by color component into blocks each having a predetermined number of pixels, sequentially perform orthogonal transformation for each divided block, and then quantize the transformation output by a quantization means, and then, This quantized output is given to variable length encoding means to be variable length encoded, and each block is determined based on the total code amount target value to be stored and the statistics for each block. When performing variable length encoding while controlling the code amount for each block according to the allocated code amount for each block, the quantization is performed using a provisional quantization width, and the quantized output obtained thereby is encoded. The first step is to check the amount of code for the entire image and the amount of code for each block required for processing and optimization, and calculate the amount of code required for optimization from the amount of code for the entire image obtained in this first step. a second step of predicting the quantization width; and predicting the code amount for each color component when quantized with the predicted optimal quantization width, and determining the total code amount target based on the predicted code amount for each color component. a third step of proportionally allocating the values and determining the amount of code for each color component, and allocating this proportionally based on the amount of code of each block to obtain a standard allocated amount of code for each block; a fourth step of performing the quantization for each block using the quantization width; and quantizing the quantization output for each block in this fourth step to a variable length within the allocated code amount for the block. a fifth step of encoding, and a block to be currently subjected to variable length encoding processing in the fifth step;
The difference between the code amount by variable length coding in the block processed immediately before and the allocated code amount in this block is calculated, and this is added as a carryover amount to the standard allocated code amount in the block to be currently processed. and a sixth step of setting the amount of code to be allocated to the block to be currently processed. (11) Divide the color image data separated by color component into blocks each having a predetermined number of pixels, sequentially perform orthogonal transformation for each divided block, and then quantize the transform output by a quantization means, and then, This quantized output is given to variable length encoding means to be variable length encoded, and when encoding data, each block is determined from the total code amount target value determined in advance and the statistics for each block. When performing variable length encoding while controlling the code amount for each block according to the allocated code amount, the quantization is performed using a provisional quantization width, and the resulting quantized output is encoded. The first step is to check the amount of code for the entire image, the amount of code for each color component, and the amount of code for each block for each color component necessary for optimization, and the code for the entire image obtained in this first step. The second step is to predict the quantization width required for optimization for each color component from the amount of code and the amount of code for each color component, and to predict the amount of code for each color component when quantized with the predicted optimal quantization width. , the total code amount target value is proportionally distributed based on the predicted code amount for each color component, the code amount for each color component is determined, and this is proportionally allocated based on the code amount for each block, and each color component is a third step of determining a standard allocated code amount for each block; and a fourth step of performing the quantization for each block for each color component using the predicted quantization width for each color component. a fifth step of variable-length encoding the quantized output for each block for each color component in the fourth step within a range that falls within the allocated code amount for the block; and a current variable-length code in the fifth step. Find the difference between the code amount due to variable length coding in the block processed immediately before the block to be processed and the allocated code amount in this block, and use this as the carryover amount for the block to be processed currently. a sixth step of adding the standard allocated code amount to the allocated code amount for the block to be currently processed. (12) Divide the color image data separated by color component into blocks each having a predetermined number of pixels, perform orthogonal transformation on each divided block in order to separate it into a DC component and an AC component, and use the orthogonal transformation output as Quantization is performed by the quantization means, and then this quantized output is given to the variable length encoding means for variable length encoding. When encoding data, the total code amount target value determined in advance and for each block are When performing variable length encoding while controlling the code amount for each block according to the allocated code amount for each block determined from the statistics of , the quantization is performed using a provisional quantization width, thereby The first step is to perform encoding processing on the obtained quantized output and check the amount of code for the entire image and the amount of code for each block required for optimization, and the code for the entire image obtained in this first step. a second step of predicting the quantization width necessary for optimization from the amount; and predicting the amount of code for each color component when quantized with the predicted optimal quantization width, and calculating the amount of code for each predicted color component. Proportional allocation of the total code amount target value based on the code amount to determine the code amount for each color component, and proportional allocation based on the code amount of each block to determine the standard allocated code amount for each block. a fourth step of performing the quantization for each block using the predicted quantization width; and assigning the quantization output for each block in this fourth step to that block. Variable-length coding is performed within the code amount, and among the image data for each block obtained by the orthogonal transformation, for the DC component, the difference with the DC component in the previous processing block is obtained. , a fifth step in which this difference is variable-length coded, and after the coding process is completed for the DC component, the AC component is variable-length coded; and the current variable-length coding process in the fifth step is performed. Find the difference between the code amount by variable-length encoding in the block processed immediately before and the allocated code amount in this block, and use this as the carryover amount to calculate the reference allocation in the block to be currently processed. and a sixth step of adding the code amount to the code amount to be allocated to the block to be currently processed. (13) Divide the color image data separated by color component into blocks each having a predetermined number of pixels, sequentially perform orthogonal transformation for each divided block, and then quantize the transformation output by a quantization means, and then, This quantized output is given to variable length encoding means to be variable length encoded, and when encoding data, each block is determined from the total code amount target value determined in advance and the statistics for each block. When performing variable length encoding while controlling the code amount for each block according to the allocated code amount, the quantization is performed using a provisional quantization width, and the resulting quantized output is encoded. The first step is to check the amount of code for the entire image, the amount of code for each color component, and the amount of code for each block for each color component necessary for optimization, and the code for the entire image obtained in this first step. The second step is to predict the quantization width required for optimization for each color component from the amount of code and the amount of code for each color component, and to predict the amount of code for each color component when quantized with the predicted optimal quantization width. , the total code amount target value is distributed proportionally based on the predicted code amount for each color component, the code amount for each color component is determined, and this is proportionally allocated based on the code amount for each block for each color component. , a third step of determining a standard allocated code amount for each block for each color component, and a fourth step for performing the quantization for each block for each color component using the predicted quantization width for each color component. step, the quantized output for each block in this fourth step is variable-length encoded within the allocated code amount for that block, and the quantization output for each block obtained by the orthogonal transformation is For the DC component of the image data, the difference between the DC component and the DC component in the previous processing block is obtained, this difference is variable-length encoded, and after the encoding process is completed for the DC component, the AC component is variable-length coded. a fifth step of performing encoding processing; and the amount of code obtained by variable length encoding in the block processed immediately before the block to be currently subjected to variable length encoding processing in the fifth step, and the amount of allocated code for this block. and a sixth step of calculating the difference between the two and adding this as a carryover amount to the reference allocated code amount for the block to be currently processed, and setting it as the allocated code amount for the block to be currently processed. An image data encoding method. (14) Divide the color image data separated by color component into blocks each having a predetermined number of pixels, perform orthogonal transformation on each divided block in order to convert it into data by frequency component, and use this orthogonal transformation to The obtained data for each frequency component is quantized sequentially from the low frequency component by the quantization means, and then this quantized output is given to the variable length encoding means to be variable length encoded. When performing variable length encoding while controlling the code amount for each block according to the total code amount target value to be stored and the allocated code amount for each block determined from the statistics for each block, the above quantization is temporarily performed. The first step is to perform encoding processing on the quantized output obtained by this method, and to check the amount of code for the entire image and the amount of code for each block required for optimization. , a second step of predicting the quantization width necessary for optimization from the code amount of the entire image obtained in the first step; and a code for each color component when quantized with the predicted optimal quantization width. The total code amount target value is allocated proportionally based on the predicted code amount for each color component, the code amount for each color component is determined, and this is allocated proportionally based on the code amount for each block. a third step of calculating a reference allocated code amount for each block; a fourth step of performing the quantization for each block using the predicted quantization width; The quantized output for each block is variable-length encoded within the allocated code amount for that block, and among the image data for each block obtained by the orthogonal transformation, the DC component is A fifth step of obtaining a difference between the DC component and the DC component in the processing block, variable-length encoding the difference, and after completing the encoding process for the DC component, performing variable-length encoding process for the AC component; In the fifth step, the difference between the code amount due to variable length coding of the block processed immediately before the block to be subjected to variable length coding processing and the allocated code amount for this block is calculated, and this is carried over as the amount of code. and a sixth step of adding the standard allocated code amount in the block to be currently processed as the allocated code amount in the block to be currently processed as the allocated code amount in the fifth step. A method for encoding image data, characterized in that variable length encoding processing is applied only to the alternating current component within a range within a certain amount. (15) Divide the color image data separated by color component into blocks each having a predetermined number of pixels, perform orthogonal transformation on each divided block in order to convert it into data by frequency component, and use this orthogonal transformation to The obtained data for each frequency component is quantized sequentially from the low frequency component by the quantization means, and then this quantized output is given to the variable length encoding means to be variable length encoded. When performing variable length encoding while controlling the code amount for each block according to the total code amount target value to be stored and the allocated code amount for each block determined from the statistics for each block, the above quantization is temporarily performed. The resulting quantization output is encoded, and the code amount for the entire image necessary for optimization, the code amount for each color component, and the code amount for each block for each color component are A first step of examining the amount of code, and a second step of predicting the quantization width required for optimization for each color component from the amount of code for the entire image obtained in the first step and the amount of code for each color component. , predict the code amount for each color component when quantized with the predicted optimal quantization width, and proportionally distribute the total code amount target value based on the predicted code amount for each color component, and calculate the code amount for each color component. a third step of determining a standard allocated code amount for each block for each color component by allocating it proportionally based on the code amount for each block for each color component; a fourth step of performing the quantization for each block for each color component using the quantization width; and a range in which the quantization output for each block in this fourth step falls within the allocated code amount for the block. Among the image data for each block obtained by the orthogonal transformation, the difference between the DC component and the DC component in the previous processing block is obtained, and this difference is variable-length coded. After the DC component is long encoded and the encoding process is completed, the alternating current component is subjected to variable length encoding. The difference between the code amount by variable length coding in the block processed immediately before and the allocated code amount in this block is calculated, and this is added as a carryover amount to the standard allocated code amount in the block to be currently processed. , a sixth step in which the code amount is allocated to the block to be currently processed, and variable length encoding processing within the range of the allocated code amount in the fifth step is applied only to the alternating current component. An image data encoding method characterized by: (16) The image data encoding method according to any one of claims (10) to (15), wherein the provisional quantization width uses a predetermined value depending on the total amount of code per screen. (17) The code of image data according to any one of claims (10) to (15), characterized in that the provisional quantization width uses a predetermined value according to the selected total code amount per screen. method.